Jana2006 Cookbook

Version October 2017

Solved examples on various topics covered by crystallographic computing
system Jana2006. The cookbook cumulates examples from all Jana workshops.
The program Jana2006, the latest cookbook and data for examples are available
in jana.fzu.cz.

Authors

Michal Dušek, Václav Petříček, Lukáš Palatinus, Jan Rohlíček, Karla Fejfarová,
Václav Eigner, Monika Kučeráková, Morgane Poupon and Margarida Henriques

Institute of Physics ASCR, v.v.i., Na Slovance 2, Prague, Czech Republic

In case of troubles with solution of examples or installation of programs please

contact Václav Petříček ([email protected]) or Michal Dušek ([email protected]).
2 Jana2006 Cookbook, version October 2017
 Contents
Example 1.1: Zn ........................................................................................ Page 11
Application of Jana2006 to simple 3d structure. Creating jobname, reading input files,
determination of symmetry, solving structure with charge flipping, editing of atomic
parameters, refinement, hydrogen assignment.
Example 2.1: PbSO4 .................................................................................. Page 21
Simple structure from powder data
Example 2.2: Y2O3 ..................................................................................... Page 27
Solution of profile asymmetry in powder data
Example 2.3.1: MagCag ............................................................................ Page 33
Solution of simple organometallic structure from powder data
Example 2.3.2: MagCag ............................................................................ Page 41
Preparation of the rigid body template in M45
Example 2.3.3: MagCag ............................................................................ Page 47
Solution of simple organometallic structure from powder data, using the rigid body
template from M45
Example 2.4: PFPhenyl ............................................................................. Page 59
Refinement of moderately sized organometallic structure from powder data with
help of geometry constrains
Example 2.5: TwoPh ................................................................................. Page 73
Example of multiphase powder refinement
Example 3.1: AD3 ..................................................................................... Page 83
Simple structure with pseudo-merohedric twinning and unequal twin volume
fractions. Finding twinning matrix by Symmetry wizard. Creating publication CIF
Example 3.2: PyNinit ................................................................................ Page 91
Simple structure with non-merohedric twinning. Handling twin overlap in Jana2006.
Example 3.3: CsLiSO4 ................................................................................ Page 99
Simple structure with pseudo-merohedric 3-fold twinning. Finding twinning matrix
from group→subgroup transformation. Transformation to four times larger
reciprocal cell.
Example 3.4: Ca2Fe2O5 ............................................................................ Page 109
HT structural phase transition, coexistence of two phases in single crystal with strong
overlaps.
Example 4.1: PtCu .................................................................................. Page 117
Disordered structure.
Example 4.2: NaNiFe .............................................................................. Page 127
Mixed sites.

3 Jana2006 Cookbook, version October 2017

Example 5.1: YPO ................................................................................... Page 133
Simple incommensurately modulated structure with crenel
Example 5.2: Na2CO3 .............................................................................. Page 147
Simple incommensurately modulated structure
Example 5.3.1: Cr2P2O7 data ................................................................... Page 173
Processing of the area detector data by Crysalis to prepare input files for Example
5.3.2
Example 5.3.2: Cr2P2O7 ........................................................................... Page 185
Incommensurately modulated structure with crenel functions and with disorder close
to the points of discontinuity
Example 5.4: Fresnoite ........................................................................... Page 203
Incommensurately modulated (3+2)-dimensional structure
Example 5.5.1: Melilite data ................................................................... Page 213
Processing of the area detector data by Crysalis to prepare input files for Example
5.5.2
Example 5.5.2: Melilite ........................................................................... Page 227
Incommensurately modulated (3+2)-dimensional structure
Example 5.6: Iodomol ............................................................................. Page 237
A molecular modulated structure from single crystal data – molecular modulations
and tests of positional vs occupational modulation
Example 5.7: CaCNi................................................................................. Page 253
A simple molecular modulated structure from single crystal data with crenel
functions
Example 6.1: KSm ................................................................................... Page 267
Modulated structure solved from powder data. Basic steps with powder refinement,
mixed sites in 4d.
Example 7.1: Ephedrine .......................................................................... Page 279
Commensurately modulated structure with merohedric twinning
Example 7.2: CrPOcom ........................................................................... Page 285
Solution of the low temperature commensurate phase of Cr2P2O7.
Example 7.3: PhenanTin ......................................................................... Page 297
Commensurate and supercell description of a five-fold superstructure.
Example 8.1: TerBut ............................................................................... Page 315
Application of rigid body groups for disordered parts of crystal structure
Example 8.2: Dinitros ............................................................................. Page 329
Application of rigid body groups for disordered parts of crystal structure
Example 9.1: L9A .................................................................................... Page 339
Getting familiar with the output of Superflip. Some useful tricks.

4 Jana2006 Cookbook, version October 2017

Example 9.2: flo19 .................................................................................. Page 349
Determining symmetry with Superflip.
Example 9.3: MnVPO .............................................................................. Page 361
A difficult case of symmetry.
Example 10.1: AgSbS .............................................................................. Page 369
Refinement of anharmonic ADP.
Example 10.2: Oxa .................................................................................. Page 381
Multipole refinement.
Example 11.1: Bezcox ............................................................................. Page 389
Composite structure.
Example 12.1: Ba5Co5 ............................................................................. Page 401
Simple antiferromagnetic magnetic structure (k = 0).
Example 12.2: Ba6Co6 ............................................................................. Page 413
Commensurate magnetic structure, k=(0,0,1/2).
Example 12.3: PrSrMnO.......................................................................... Page 421
Commensurate magnetic structure, k=(1,0,0)
Example 12.4: MnWO4 ........................................................................... Page 431
Incommensurate magnetic structure, k=(-0.209 0.5 0.453)
Example 12.5: HoNi ................................................................................ Page 439
Magnetic structure refined from single crystal data - k=(0,0,1)
Example 12.6: ISODISTORT ..................................................................... Page 445
Practical introduction to Jana2006-ISODISTORTcommunication
Example 12.7: DyMn6Ge6 ...................................................................... Page 453
Incommensurate magnetic structure – k=(0,0,0.1651); k=(0,0,0)
Example 12.8: Cr2WO6 .......................................................................... Page 471
Commensurate magnetic structure – k=(0,0,0)
Example 13.1: TTB .................................................................................. Page 481
Solution of a tetragonal-tungsten-bronze type structure from EDT (electron
diffraction tomography) data. Complete procedure from data reduction to structure
refinement
Example 13.2: Kaňkite ............................................................................ Page 491
Solution of kaňkite – a hydrated Fe-bearing arsenate – from electron diffraction data
Example 13.3: Orthopyroxene ................................................................ Page 498
Dynamical structure refinement of occupancies
Example 13.4: Silicon .............................................................................. Page 506
Solution and refinement of silicon structure from three zone-axis patterns
Example 13.5: Orthopyroxene from EDT................................................. Page 525

5 Jana2006 Cookbook, version October 2017

A complete procedure from structure solution to dynamical structure refinement
from electron diffraction tomography data
Example 20.1: CeAsSe............................................................................. Page 535
Structure import: merohedric twin with extinction from SHELX
Example 20.2: HfPdGe ............................................................................ Page 537
Structure import: Six-fold twin with reflections in HKLF5 format from SHELX
References ............................................................................................. Page 539
List of references related to methods used in Jana2006, including external programs
for charge flipping and electron diffraction
Appendix 1.1: File format PRF ................................................................ Page 541
PRF file is used for powder refinement. It contains Bragg positions and profile points

6 Jana2006 Cookbook, version October 2017

Installation notes
For the workshop, following programs are necessary:

• Jana2006 for most tasks

• Diamond for plotting crystal structures
• Superflip (part of Jana2006 distribution) for solution of phase problem
• Vesta for visualization of electron density with isosurfaces and for magnetic structures
• PETS for data processing of electron diffraction patterns
• ImageJ for viewing graphical output from PETS

Diamond can be replaced by any other plotting software capable to read CIF in command line. For the
workshop we recommend Diamond because the connection with Jana2006 is very well tested.

Installation of Diamond
Install demo version of Diamond from the directory Setup\Diamond installation in the flash disk.
Copy the licence file “yourlicence.lic” to the same directory where Diamond.exe has been installed.
Usually it will be c:\Program files\Diamond 4
This time limited license has been kindly provided by Crystal Impact for this workshop.
Note: Diamond started from Jana2006 may behave differently than when started separately. For
instance, it doesn’t offer automatic wizards.

Installation of Vesta
Unpack the archive “Vesta_2.1.2.zip” from the directory Setup\Vesta to a directory of your choice

Installation of Jana2006
• Don’t use your “old” preinstalled
Execute janainst.msi from the directory versions of Jana2006
Setup\Jana2006 and follow instructions.
The installation creates shortcut to Jana2006 in the • Don’t use flash disk for calculation.
desktop and in Start → Programs Copy the example files to hard disk.
Start Jana2006, go to “Tools → About Jana2006” and
verify the version string. • Don’t install Jana2006 to desktop or
The version date should be close to the workshop date flash disk

Configuration of Jana2006
Start Jana2006
Go to “Tools → Preferences”, adjust size of the window (for instance 60% of the screen height) and font
height (about 11 points)
Go to “Tools → Programs”; focus with mouse the textbox “Graphic viewer”; press “Browse” and localize
Diamond.exe (usually in Program Files x86)
Focus "3d visualization of maps"; press “Browse” and localize VESTA.exe
(VESTA is needed for examples 9.x, 10.x, 12.x and 13.x)
For using a different text editor than Notepad define its pathname in the textbox “Editor name”.

7 Jana2006 Cookbook, version October 2017

Installation of ImageJ (only for electron diffraction)
Install ImageJ from the installer in the directory Setup\ImageJ to a directory of your choice. Use ij148-
jdk6-64bit-setup.exe, if you have a 64-bit system. Use ij148-jdk6-setup.exe for a 32-bit system, or if you
are not sure.

Installation of PETS (only for electron diffraction)

Copy the file pets.exe from the directory Setup\PETS to a directory of your choice.

Add ImageJ and PETS in the system variable PATH:

Under Windows XP:
Go to “Start → Settings → Control Panel”
Double click “System”
Go to “Advanced → Environment variables”
In the list “System variables” find and double click “Path”
In the small window that opens, add the full paths to the files ImageJ.exe and pets.exe (without
the file names), separated by a semicolon
Under Windows 7:
Right click on the Computer icon on your Desktop and choose “Properties → Advanced system settings
→ Environment variables
In the list “System variables” find and double click “Path”
In the small window that opens, add the full paths to the files ImageJ.exe and pets.exe (without
the file names), separated by a semicolon

Launching PETS:
PETS must be called with the input file as a command-line argument. There are several ways of launching
PETS:
Method 1 – the most convenient:
Locate the file .pts (input to PETS) on your system
Right-click to open context menu
“Open → Select program from a list → Browse…”
Locate the file pets.exe on your system; Open
Make sure the checkbox next to “Always use the selected program to open this kind of file” is checked
OK
PETS starts. Next time you want to run PETS, it is sufficient to double click the icon of the pts file you
want to run. Unfortunately, this method fails on some systems, notably if the extension pts is already
associated with another software.
Method 2:
Place the shortcut to PETS on Desktop:
Locate the file pets.exe on your system
Select it and copy it to clipboard (Ctrl+C)
Place cursor on the place on the desktop where you want to place the icon of PETS
Right-click → Paste shortcut
Now you can run PETS by drag-and-dropping the file .pts on the icon on the desktop. Unfortunately, even
this method does not work on all systems.
Method 3 – works always:
Open DOS terminal (Start → All programs → Accessories → Command Prompt)
Move to the directory with the file .pts you want to use (use command cd)
Run PETS by typing “pets filename.pts”, where filename.pts is the file you want to use.

8 Jana2006 Cookbook, version October 2017

Upgrade of Jana2006 installed from janainst.msi (new way of installation)
Execute janainst.msi and follow instructions. Make sure the installation directory is the same like for the
previous version.
The installation wizard will reinstall Jana2006. For the same version it will offer “Modify – Repair –
Remove”.

Upgrade of Jana2006 installed from janainst.exe (old way of installation)

Previously Jana2006 was installed from the self extracting archive janainst.exe. The user had to define
“by hand” the system environment variable JANA2006DIR. The installation from janainst.msi in some
cases cannot reinstall Jana2006 installed from janainst.exe.
Start Jana2006 (old installation)
Go to “Tools → About Jana2006” and note the version string.
Start janainst.msi and follow instructions. Make sure the installation directory is the same like for the
old version.
Start Jana2006 (new installation)
Go to “Tools → About Jana2006” and verify the version string is new.
Troubleshooting (for upgrade of Jana2006 installed from janainst.exe)
1. Installation from janainst.msi was running without warnings but the version string of
Jana2006 remains old
Start janainst.msi again. It will offer “Modify – Repair – Remove”. Used “Remove” and uninstall old
version of Jana2006.
Start janainst.msi (the third time) and follow instructions. Make sure the installation directory is the
same like for the old version.
Start Jana2006 (new installation)
Go to “Tools → About Jana2006” and verify the version string is new.
2. Another issues or when method 1 fails
Localize installation directory of Jana2006 and make backup copy of jana2006.ini and jana2006.hst
Delete installation directory of Jana2006
Go to “Start → Settings → Control Panel → System”
Go to page “Advanced”
Press “Environment variables”
Delete all instances of JANA2006DIR in User variables as well as in System variables
Start janainst.msi and follow instructions.
Return files jana2006.hst and jana2006.ini to the installation directory (where jana2006.exe is located)
Start Jana2006 (new installation)
Go to “Tools → About Jana2006” and verify the version string is new.

9 Jana2006 Cookbook, version October 2017

10 Jana2006 Cookbook, version October 2017
Example 1.1: Zn
Application of Jana2006 to a simple 3d structure. Creating jobname, reading
input files, determination of symmetry, solving structure with charge flipping,
editing of atomic parameters, refinement, hydrogen assignment.

Revised: 25 September 2017

Chemical formula: (NH3CH2CH2NH3)2ZnCl6

Single crystal data measured with Oxford Diffraction four-circle diffractometer
Input files: Zn.hkl, Zn_red.sum
Frame scaling, absorption correction: done with software of diffractometer

1. Creating new jobname

Start Jana2006
“File → Structure → New” opens a file manager
Left pane: locate the directory with input files
Right pane: double-click Zn
Right pane detects possible Jana files and shows one jobname for each group of files

2. Import Wizard
Select “Known diffractometer formats”; NEXT
Select “Oxford Diffraction → CCD”; NEXT
Select “Input from “sum” file”; OK
Leave all settings unchanged; NEXT
Leave all settings unchanged; NEXT
The program reads 24240 reflections from hkl file
For absorption correction select “None or done before importing”; NEXT
FINISH

3. Data Repository
Each line in data repository window corresponds to one data set converted to Jana format.
In our case, we use only one data set. Data repository is saved in file Zn.m95.
OK; YES to accept the data set

4. Symmetry Wizard
[On the screen: “Tolerances for crystal system recognition”.]
Symmetry Wizard can be started separately by “File → Reflection file → Make space group
test.
Leave all settings default; NEXT
[On the screen: Select Laue symmetry]
Select Orthorhombic Laue symmetry; NEXT
Select primitive unit cell; NEXT
[On the screen: Select space group]
The window shows possible space groups.
“Obs/all” column shows for each space group number of reflections, which should be

11 Jana2006 Cookbook, version October 2017

 systematically extinct for the selected space group. For instance, Pbnm has
“Obs/all”=2/1677, which means 1677 reflections in the reflection file should be
systematically absent, i.e. with intensity below the 3σ limit, but two of them have anyway
intensity above the 3σ limit and therefore violate the rules for systematic absences.
“Ave(I/sig(I)” lists average ratio I/σ (I) for columns “Obs/all”. For Pbnm this means that 1677
theoretically absent reflections have the average I/σ (I) = 0.153 and the two of them, which
violate the rules for systematic absences have the average I/σ (I) = 4.783.
The limits for the average ratio I/σ (I) have been defined previously in “Tolerances” window.
Space groups exceeding the limits are not offered.
For selecting the proper space group, we are looking for the cases with high number of
unobserved systematically absent reflections having at the same time reasonably low
average I/σ (I) for reflections violating the extinction rules. In our example, Pbnm is more
likely than Pbmm because number of experimentally confirmed absent reflections is much
larger.
The list of the strongest reflections contradicting the selected space group can be displayed
by “Details” button.
Jana2006 does not test data for centre of symmetry. At this point centrosymmetric space
group should be preferred unless we would have a strong reason for non-centrosymmetric
space group. For this reason, we select Pbnm and not Pbn21. Centrosymmetry is usually
confirmed or denied in following steps by the symmetry analysis by Superflip (see Step 6).
Select Pbnm; NEXT
Accept the space group in the standard setting, i.e. like Pnma; FINISH
Symmetry is saved in file Zn.M50.

5. Creating refinement reflection file

In this step program creates file Zn.M90 containing the data set merged by symmetry and
with discarded forbidden reflections. M90 will be used for refinement.
NEXT to confirm threshold 3sigma; OK; OK (to confirm Information windows)
[On the screen: settings for merging of reflections]
Check “Display graph sig(Icount)/sig(Istat)”; NEXT
[On the screen: graphs of sigma]

12 Jana2006 Cookbook, version October 2017

 The vertical axis shows sigma of reflections calculated from merging statistics.
The horizontal axis for black circles: sigma as obtained from the diffractometer file
The horizontal axis for blue circles: sigma obtained from the diffractometer file and
corrected by the “instability factor” calculated such that the values from the diffractometer
file are fitted to the values from merging statistics.
For highly redundant data sets with reliable merging statistics, it is recommended to use
such “instability factor” for the weighting scheme for refinement (done at the end of this
example).
QUIT
The program reports Rint 3.97% for 1259 observed reflections merged from 12466 observed
reflections.
Select "Accept the new DatBlock and calculate coverage"; FINISH

6. Structure Solution Wizard

[On the screen: Structure solution]
Structure solution wizard can be executed separately through “Run → Structure solution” or
by double clicking the icon “Structure solution” in the main window.
Jana2006 does not contain solution procedures, it calls external programs. The program
used in most our examples is Superflip (using the charge flipping algorithm) because it is
distributed together with jana2006. For standard structures, good alternative is SHELXT but
it should not be distributed without registration so users should connect it individually.

13 Jana2006 Cookbook, version October 2017

 In “Formula” textbox type list of chemical elements: Zn Cl C N H
Superflip as well as SHELXT does not need exact information about chemical elements
because it is required only for analysis of the resulting electron density. Other solution
methods like SIR may need the correct formula and number of formula units.
Select “Use Superflip”; leave other settings default; press “Run solution”
[On the screen: window of Superflip replaced after reaching the convergence with listing of
Superflip]
Superflip converges (after noise suppression) with R value ~15%. For charge flipping, this is
very good result. The space group suggested from the symmetry search in the electron
density map is Pnma, i.e. the same like obtained previously in Jana2006. The low agreement
factors at the bottom of the listing indicate good fit of the user-defined symmetry elements
(here Pnma) with the resulting electron density map.
Press CLOSE to leave the listing

7. Verification of the structure solution from the structure solution wizard

[On the screen: Structure solution]
Press “Draw structure”
Press "Draw + continue"
[On the screen: window of Diamond]
Start “Build → Get molecules” or use on the bottom toolbar
Rotate the molecule ( and on the bottom toolbar) to get an optimal view
Move mouse pointer over the terminal carbon atoms. Diamond will show labels of the
atoms. Note labels (probably C1,C2) of the terminal carbons.
By comparison with the chemical diagram, it is clear that the terminal carbon atoms are in
fact nitrogen. We have to note their labels in order to change them later on to nitrogen.
[On the screen: plot of the structure in Diamond]

14 Jana2006 Cookbook, version October 2017

 Quit Diamond (without saving changes)
[On the screen: Structure solution]
Press “Draw 3d map”
This button starts Vesta (requires path to Vesta set in Tools → Programs), which plots the
electron density produced by Superflip together with atomic positions found by Jana2006

Quit VESTA without saving the plot

Press “Accept last solution”. This action will also close the structure solution wizard

8. Assignment of correct chemical types

[On the screen: basic Jana window]
Start EditM50; go to page Composition; enter 4 for “Formula units”, press “Formula from
M40”
It displays formula Zn Cl6 C8 (for Z=4). However, the expected composition (without
hydrogens) is Zn Cl6 C4 N4. This means that some nitrogens have been falsely assigned as
carbons.The expected chemical formula also suggests that nitrogens should be at the
terminal positions of the organic part.
[On the screen: Information about formula]
OK; YES for updating the formula.
Numbers of elements in the formula are not important for refinement. This information is
useful only for interpretation of difference Fourier maps and for creating the final CIF. The
chemical types in M40 are defined by their sequence numbers in the chemical formula.
Nevertheless, the order of atoms in the formula can be arbitrarily changed if it is done

15 Jana2006 Cookbook, version October 2017

 through EditM50. Also new chemical elements can be included in arbitrary position of the
formula.
OK; YES to rewrite M50

Start “Edit atoms”

[On the screen: list of atoms]
Select (with Ctrl-left mouse button) the previously noted terminal carbon atoms
Plotting of the structure can be repeated by “Plot structure” icon
Start “Action → Edit/define”
Choose “Type” nitrogen; OK
[On the screen: list of atoms. The labels are still the original ones but the chemical type of
two selected atoms is nitrogen]
Press “Select All”
Press “Action” or right-click on one of the selected atoms
Choose “Rename selected atoms to atom_type + number”
[On the screen: list of atoms with labels corresponding to chemical types]
You should see one zinc, four chlorines, two nitrogens and two carbons.
Press “Select All”
Press “Action” or right-click on one of the selected atoms
Choose “Make symmetrically contiguous motifs”
This makes happy editors of Acta E and it is also necessary for using rigid body (later in this
Cookbook). The distances between atoms in a motif, typically molecules, can be calculated
without applying any symmetry operation except identity.
OK; YES to save the changes

9. Refinement
[On the screen: basic window of Jana]
Right-click the icon of Refine.
[On the screen: refinement options]
Define 100 of cycles; OK
Choose “YES+START”
Refinement converges with R value about 6%
Start “Edit atoms”
[On the screen: list of atoms]
Press “Select all” and “Action → Edit/define”
Choose “harmonic” for ADP parameters. This defines anisotropic temperature parameters
for all atoms. OK, OK, Yes to save changes
Double-click the icon Refine.
Refinement converges with R value about 4 % for 73 refined parameters.

10. Adding of hydrogen atoms

[On the screen: basic window of Jana]
Start “Edit atoms”
Select carbons and nitrogens
Start “Action → Adding of hydrogen atoms”
[On the screen: Adding of hydrogen atoms dialogue]
Leave all settings default; OK
The procedure adds automatically C-H hydrogen atoms, while for adding hydrogens to the
16 Jana2006 Cookbook, version October 2017
 terminal NH3 groups it goes to the manual mode.
[On the screen: Adding of hydrogen atoms for N1]
For the type of coordination select “Tetrahedral”, change number of neighbors to 1
Using tetrahedral instead of trigonal geometry around nitrogen follows from the chemistry.
The terminal groups are expected as NH3 because of the charge balance (unlike NH2, they
have positive charge).
The only neighbor of N1 is C1 (or C2 depending how the maxima were assigned after
running Superflip). You can verify this by focusing the textbox with C1 (or C2) and pressing
“Select neighbors”
Define names of hydrogens, for instance H1N1, H2N1 and H3N1
Start “Locate position in the map”
[On the screen: difference Fourier map showing the hydrogen maxima]
Press “Optimal” and “Apply”
Locating position in the map is not necessary. In most cases, the rotation of NH3 or CH3
group can be refined correctly from the initial positions. However, in our case we want to
confirm that there are really NH3 groups and not NH2.
[On the screen: Adding of hydrogen atoms to N1]
“Apply → Go to next”
Repeat the procedure for N2
[On the screen: list of atoms]
Total number of atoms should be 19 (can be verified with Select All).
The ideal geometry of hydrogen atoms is kept by commands, which can be viewed or edited
in “Refinement commands → Various → Keep commands”.
OK to close Atoms-Edit; Yes to save changes
“Plot structure” to check results of hydrogen assignment
[On the screen: window of Diamond]
Start “Build → Connectivity” or press button
[On the screen: Connectivity dialogue]
Press "Evaluate" to see which bonds will be used by Diamond
Plot the structure with
Use at the bottom toolbar of Diamond to see ADP ellipsoids.
Quit Diamond

[On the screen: basic window of Jana2006]

Double-click Refine and refine the final structure
Refinement converges with R factor 2.5%, goodness of fit 1.9. Number of refined parameters
is still 73because the positions of hydrogen atoms are defined geometrically.

11. Extinction correction

[On the screen: basic window of Jana2006]
Open listing of Refine (by “Edit/View → View of Refine”)
Press “Go to” and select “Statistics Fo, sin(th)/lambda after refinement”
R value should decrease with increasing intensity or with lowering the diffraction angle.
Here the R value for strongest reflections as well as for the lowest angle reflections is higher
than in the neighboring shell. This suggests that extinction correction is required.

17 Jana2006 Cookbook, version October 2017

 Close the refinement listing
“Parameters → Extinction”
Select “Isotropic” “Type 1” “Gaussion”; OK
Refinement of extinction coefficient is enabled by default

12. Final refinement

[On the screen: basic window of Jana2006]
Double-click Refine and refine the final structure
Refinement converges with R factor 2.24%, goodness of fit 1.77, 74 parameters.
Check the extinction coefficient through “Parameters → Extinction”
View listing of Refine; Go to “Statistics”; check improved R value for strongest reflections.

The resulting structure

Standard refinement calculates with F2 while Jana2006 is still using F by default. In order to
get a standard result, we will re-refine the structure on F2:
Right-click the icon of Refine.
[On the screen: refinement options]
Check “Refinements on R(obs)**2”; OK; YES to start the program
Refinement converges with R factor 2.19%, goodness of fit 1.61, 74 parameters.
For good data, refinement on F and F2 usually gives very similar results.

Jana2006 does not refine the weighting scheme. The goodness of fit therefore corresponds
to the experimental sigma of reflections corrected by the instability factor. For the moment,
we are using the default instability factor 0.01. In the next step, we will use the instability
factor calculated during merging of reflections (see above):

18 Jana2006 Cookbook, version October 2017

Right-click the icon of Refine.
[On the screen: refinement options]
Check “Instability factor from reflection statistics”; OK; YES to start the program
There are almost no changes in refinement results because the calculated instability factor is
close to the default one. Its value can be seen in Edit/View → View of Reflection report
(search for the string “instability”).

19 Jana2006 Cookbook, version October 2017

20 Jana2006 Cookbook, version October 2017
Example 2.1: PbSO4
Simple structure from powder data
Revised: 23 November 2016

PbSO4
Powder data measured with laboratory diffractometer
Input files: PbSO4.mac (powder profile data)
PbSO4.txt (additional information)

1. Creating new jobname

Start Jana2006
“File → Structure → New” opens a file manager
Left pane: locate directory with input files
Right pane: double-click PbSO4

2. Import Wizard
Select “Various CW formats”, NEXT
CW stands for constant wavelength
Select “MAC format” and “Another/unknown method”; NEXT
File name is automatically set to”PbSO4.mac”.
Fill cell parameters 8.48 5.4 6.96 90 90 90
For wavelength type first select Kalpha1/Kalpha2 doublet and then select X-ray tube Cu
Select “Parallel setting”, “Glancing angle” 13.28815, “Perfectness 0.5”; NEXT; FINISH
Accept the data in Data repository
Select "Yes, I would like to continue with the wizard"; OK
At this point Jana2006 offers wizard for determining powder profile parameters, space
group and structure. We shall use the wizard but all its tools could be also started
separately.

3. Refinement of the powder profile

[On the screen: Refinement of the powder profile by the le Bail algorithm]
“Show powder profile”
Profile viewer shows experimental powder profile
Leave profile viewer
“Edit refinement commands”; set 0 cycles; leave other settings default; OK
“Run Refine”
“Show powder profile”
Now the profile viewer shows experimental and calculated powder profile.
The calculated profile is based on default profile parameters.
The calculated profile is stored in PbSO4.prf
Leave profile viewer
“Edit profile parameters”
[On the screen: Powder options]
In “Cell” page select refinement of a, b and c
In “Profile” page select Cutoff to 12*FWHM; activate refinement of GW;
In “Corrections” page activate refinement of “shift”;
21 Jana2006 Cookbook, version October 2017
In “Corrections” page select “Legendre polynomials” with 5 terms for background
calculation
Refinement of background terms is enabled by default
Leave “Edit profile parameters”, save changes
“Edit refinement commands”; set 10 refinement cycles; OK
“Run Refine”
After 10 refinement cycles Rp~11.4%.
“Show powder profile”
In Powder profile viewer press “Default” and move close to theta 20°

The observed peaks are systematically broader which may indicate that Lorentzian
broadening is to be combined with the Gaussian profiles.
“Edit profile parameters”
[On the screen: Powder options]
In page “Profile” select Pseudo-Voigt and activate refinement of LY
“Run Refine”
After10 refinement cycles Rp~8%
“Show powder profile”
“Edit profile parameters”
[On the screen: Powder options]
In page “Profile” activate refinement of LX.
“Run Refine”
After 10 refinement cycles Rp~6.2%
“Show powder profile”

22 Jana2006 Cookbook, version October 2017

 The observed peaks at low angles still have some asymmetry. (Use “X+” and “X-“ buttons to
see it better)
“Edit profile parameters”
[On the screen: Powder options]
In page “Asymmetry” select the Berar-Baldinozzi correction and activate refinement of
asym1, asym2, asym3, asym4;
“Run Refine”
After 10 refinement cycles Rp~5.4%
FINISH; Yes to close the refinement tool; Next

4. Space group determination

[On the screen: Tolerances for crystal system recognition]
Use default; NEXT
[On the screen: Select Laue symmetry]
Select the highest one – orthorhombic; NEXT
[On the screen: Select cell centering]
The table shows Rp values calculated for various centering vectors. For instance, when we
introduce A centering and discard corresponding extinct Bragg positions, Rp becomes about
30% . This indicates we don’t have A centering. In our case no centering is indicated.
Select P centering as it gives much better profile fit; NEXT
[On the screen: Information about progress of space group tests. It takes some time…]
[On the screen: Select space group];
Among the space groups with good profile fit we are looking – in analogy with a single-
crystal experiment – for a space group with maximum number of extinct reflections
consistent with the experiment. The space groups Pnma and Pn21a have good profile fit
(Rp~5.4%) and the largest ratio of systematically extinct reflections (about 13%). We shall
select the centrosymmetric one.
Select space group Pnma; NEXT
[On the screen: Final step of the space group test];

23 Jana2006 Cookbook, version October 2017

 Accept the space group transformed into the original cell; FINISH
NO to refine profile parameters once more.

5. Structure solution
[On the screen: Structure solution]
This part of wizard can be also started separately by “Run → Solution”
Type formula Pb S O4; Formula units 4
Select “Use Superflip”
Check “Repeat Superflip until the convergence is detected”
Set “Maxcycles” to 1000

Press “Run solution” to start Superflip

Jana makes LeBail decomposition and then it starts Superflip.
Superflip confirms space group Pnma
Close listing of Superflip; “Draw structure”; “Draw+continue”.
For plotting with Diamond use button at the bottom toolbar and plot only Pb and S. In
Build → Connectivity adjust limits for S-O distances. The complete coordination of Pb and S
with .
The sulphur atom should be coordinated by four oxygen atoms. But as the structure contains
one very heavy atom it may happen that some of O atoms are missing or that the solution
contains some spurious peaks or that SO4 tetrahedron is incomplete.
If the structure model looks strange you may repeat the solution step with “Run →
Structure solution”. Superflip starts from random phases and its results can be different.

Press “Accept last solution”. This action will also close the structure solution wizard.

24 Jana2006 Cookbook, version October 2017

6. Refinement of the structure
[On the screen: basic window of Jana]
Open refinement options (by the right-click on the icon of Refine).
[On the screen: refinement options]
Uncheck “Make only profile matching”, OK, Yes+start
Wait for convergence and open the listing. (You can access the listing of Refine also by
“Edit/View → View of Refine”)
It may happen that Refinement fails. In such case use “Tools → Recover files” and repeat
Structure solution. Superflip starts from random phases and may return better result in the
next attempt.
Press “Go to” and select “Changes overview”:

At the beginning of “Changes overview” there will be probably list of spurious atoms
automatically excluded from the refinement as their ADP is larger than the allowed limit. In
this example this is O4. The excluded atoms are still present in M40 file with zero
occupation. In order to delete them:
Run “Edit atoms”, press “Select rejected”
Press the button Action and delete them; OK
Start EditM50, select page "Composition" and press "Formula from M40". It should display
"Pb S O4".

7. Completing the structure

In the case that some of oxygen atoms are still missing they can be localized from the
difference Fourier.
If your structure is complete (EditM50 displays correct formula) but you would like to try this
part, please start "Edit Atoms", delete some oxygen and repeat refinement to prepare
proper input for Fourier calculation)
Right-click on the Fourier icon opens the command editing procedure for Fourier.
Select “F(obs)-F(calc) – difference Fourier”; OK; Start the program;
Answer “Yes” to start the procedure for including atoms.
[On the screen: Inserting/replacing atoms]
The program transforms each found maximum to various symmetry equivalent positions
where it is close to some already existing atoms. In this dialogue we choose that the right
panel will show only atoms with distance to some (transformed) Fourier maximum between
1 and 3 Å. By selecting such line we select the maximum and it corresponding symmetry
transformation.

25 Jana2006 Cookbook, version October 2017

Include all missing atoms, i.e. the maxima with large enough charge and reasonable
distance to some of already existing atoms; FINISH
It may happen there are no such maxima
Verify the added maxima with plotting tool.
Verify the formula with “Editm50 → Composition → Formula from M40”
Repeat refinement (in case of changes)
Refinement should converge with Rp around 6.7%, GOF 1.8 and structure R factor around
3.5%.
“Edit/View → View of Refine”; GoTo; List of serious warnings
This list should be empty. This means for instance that ADP are positive definite.

The resulting structure

The resulting R values

26 Jana2006 Cookbook, version October 2017

Example 2.2: Y2O3
Le Bail refinement against powder data with strong asymmetry of peaks
Revised: 23 November 2016

Powder data measured with laboratory diffractometer experimental setup with Johannson-
type monochromator and transmission mode in asymmetric Guinier type arrangement.

Input files: Y2O3-icsd.cif (crystal structure from data base)

Y2O3.dat (powder diffraction data)

1. Creating new jobname

Start Jana2006
“File → Structure → New” opens a file manager
Left pane: locate directory with input files
Right pane: double-click Y2O3

2. Import procedure
Select “Structure: from CIF”; NEXT
[On the screen: Select input CIF file]
Select “Y2O3-icsd.cif”; OK
[On the screen: “CIF file does not contain any reflection block”]
Answer “Yes” to the question: “Do you want to import data from file?”
[On the screen: Specify type of the file to be imported]
Select “Powder data: various CW formats” (CW = constant wavelength); NEXT
[On the screen: Powder data from:]
Select “Riet7 formats”; NEXT
[On the screen: Complete/correct experimental parameters]
For wavelength type use “X-ray tube” Cu, deselected Kalpha1/Kalpha2 doublet and
wavelength 1.54051
Select “Parallel setting”, “Glancing angle” 13.65232 (CuKα, Ge[111]) “Perfectness 1”; NEXT;
FINISH
Accept the data in Data repository

3. Refinement of the powder profile

[On the screen: Basic Jana window]
Start "Edit profile"
[On the screen: Powder options]
In “Cell” page select refinement of the parameter a
In “Profile” page select Cutoff to 12*FWHM; activate refinement of GW
In “Corrections” page activate refinement of “shift”
In “Corrections” page select “Legendre polynomials” with 5 terms for background
calculation
Refinement of background terms is enabled by default
Leave Powder options, save changes
[On the screen: Basic Jana window]
Right click on the Refine icon;
27 Jana2006 Cookbook, version October 2017
Check “Make only profile matching”;
Keep default values for another parameters; OK;
[On the screen: Do you want to save new command?]
Select Yes+start
Refinement converges in 10 cycles to Rp~7.7%
Start “Profile viewer”
[On the screen: Jana 2006 profile viewer]
Press “X exactly” and select the 2θ interval from 19 to 35.

The observed peaks are systematically broader which can indicate that Lorentzian
broadening is to be combined with the Gaussian profiles.
Start “Edit profile”
[On the screen: Powder options]
In page “Profile” select Pseudo-Voigt and activate refinement of LY; OK
Yes to rewrite changes
Start refinement (icon Refine)
Refinement converges in 10 cycles to Rp=7.1%
Start “Edit profile”
[On the screen: Powder options]
In page “Profile” activate refinement of GU; OK
Yes to rewrite changes
Start refinement (icon Refine)
Refinement converges in 10 cycles to Rp~7.0%
Start “Profile viewer”
[On the screen: Jana 2006 profile viewer]
Press “X exactly” and select the 2θ interval from 19 to 35.

28 Jana2006 Cookbook, version October 2017

 The experimental peaks have very strong asymmetry which is not explained by the profile
model.

4. Asymmetry correction by the Simpson's model

[On the screen: Basic Jana window]
In order to save the basic refinement result for testing of all different asymmetry models a
new structure will be opened:
Start “File → Structure → Save As”
[On the screen: File manager for saving structure]
In File manager focus the right-bottom textbox and define a name for testing of the
Simpson’s model, for instance “Y2O3-Simpson”; OK
Yes to continue with the new structure
[On the screen: Basic Jana window]
Start “Edit profile”
[On the screen: Powder options]
In page “Asymmetry” select “Simpson”, use the default value and activate refinement of the
“asym1” parameter; OK;
Yes to rewrite changed files
[On the screen: Basic Jana window]
Double-click the icon Refine.
Refinement converges in 10 cycles to Rp~6.1%

Start “Profile viewer”

[On the screen: Jana 2006 profile viewer]
Press “X exactly” and select the 2θ interval from 19 to 35.

29 Jana2006 Cookbook, version October 2017

 The Simpson's model improves considerably the profile fit. However, some discrepancies due
to asymmetry are still visible.

5. Asymmetry by Berar-Baldinozzi
[On the screen: Basic Jana window]
Start “File → Structure → History”
[On the screen: History]
Select the original structure Y2O3; OK
[On the screen: Basic Jana window]
Start “File → Structure → Save As”
[On the screen: File manager for saving structure]
In File manager focus the right-bottom textbox and define a name for testing of the Berar-
Baldinozzi model i.e. “Y2O3-BB”; OK
Yes to continue with the new structure
[On the screen: Basic Jana window]
Start “Edit profile”
[On the screen: Powder options]
In page “Asymmetry” select “Berar-Baldinozzi”, use the default value and activate
refinement of the asym1, asym2, asym3 and asym4 parameters; OK;
Yes to rewrite changed files
[On the screen: Basic Jana window]
Start "Refine commands" (right-click on the icon "Refine")
In page "Basic" set 100 refinement cycles and Damping factor 0.3
OK; YES+START to save commands and start refinement.
Refinement converges after many cycles to Rp~5.8%
Start “Profile viewer”
[On the screen: Jana 2006 profile viewer]
Press “X exactly” and select the 2θ interval from 19 to 35.
30 Jana2006 Cookbook, version October 2017
 The fit is even better than for the previous method but the number of asymmetry
parameters is now four.

6. Asymmetry by axial divergence

(Finger, Cox, Jephcoat, J.Appl.Cryst.(1994). 24, 892-900)
[On the screen: Basic Jana window]
Start “File → Structure → History”
[On the screen: History]
Select the original structure Y2O3; OK
[On the screen: Basic Jana window]
Start “File → Structure → Save As”
[On the screen: File manager for saving structure]
In File manager focus the right-bottom textbox and define a name for testing of the axial
divergence correction i.e. “Y2O3-FCJ”; OK
Yes to continue with the new structure
[On the screen: Basic Jana window]
Start “Edit profile”
[On the screen: Powder options]
In page “Asymmetry” select “correction by divergence”, use the default value and activate
refinement of the HpS/L and HmS/L parameters; OK;
Yes to rewrite changed files
[On the screen: Basic Jana window]
Right-click the icon Refine
[On the screen: Refine commands]
In page "Basic" set 30 refinement cycles and Damping factor 0.5
OK; Yes+start to save the changes and start refinement
Refinement reaches after 20 cycles Rp~4.8%

31 Jana2006 Cookbook, version October 2017

Open the refinement listing. (You can access the listing of Refine also by “Edit/View → View
of Refine”).
Press “Go to” and select “Changes overview”:

The parameter HmS/L oscillates , within three times the standard uncertainty, about zero
which induces instability in the refinement. For this reason, we can change it to zero and
keep it fixed.
Start “Edit profile”
In page “Asymmetry” set HmS/L to zero and unselect it; OK;
Refinement reaches after 20 cycles Rp~4.9%
Start “Profile viewer”
[On the screen: Jana 2006 profile viewer]
Press “X exactly” and select the 2θ interval from 19 to 35.

With only
one additional free parameter the fit is significantly better than for the previous methods.

In Jana2006 there is also implemented the fundamental approach following closely the
approach described by Cheary and Coelho [J. Appl. Cryst. (1998). 31, 851-861; J. Appl. Cryst.
(1998). 31, 862-868]. This method was not applied in our example because it was developed
for Bragg-Brentano geometry.

32 Jana2006 Cookbook, version October 2017

Example 2.3.1: MagCag
Solution of simple organometallic structure from powder data
Revised: 4 October 2015
Powder data measured with diffractometer Empyrean (PANalytical) using CuKα radiation,
focusing mirror and Debye-Scherrer geometry (sample in capillary), at ambient temperature.

Input files: magcag.xrdml with powder profile

Additional information:
unit cell 9.19 21.38 8.12 90 113.78 90
composition Mg2 O20 H80 Br4 N16 C28
Expected chemical scheme:

1. Creating new jobname

Start Jana2006
“File → Structure → New” opens a file manager
Left pane: locate directory with input files
Right pane: double-click magcag

2. Import procedure
Select “Powder data – various CW formats”; NEXT
[On the screen: Powder data form]
Select "PANalytical XRDML"
For method select “Debye-Scherrer”
[On the screen: Complete/correct experimental parameters]
Fill cell parameters 9.19 21.38 8.12 90 113.78 90
Check the radiation (should be interpreted from the xrdml file as Copper with doublet):

33 Jana2006 Cookbook, version October 2017

 NEXT; FINISH; OK to close data repository; YES to accept the data set
[On the screen: question how to continue]
Select “Yes, I would like to continue with the wizard”; OK
The typical steps for refinement of powder profile will be offered in the wizard. However, all
steps can be also done separately from the basic window of Jana2006.

3. Refinement of the powder profile

[On the screen: refinement of the powder profile by the le Bail algorithm]
Press “Show powder profile” – just to see how it looks like
Press “Default” to see more details
Quit powder viewer
Press “Edit profile parameters”
Unlike for structure parameters, refinement keys of profile parameters are not set
automatically
In page “Cell” activate refinement of a,b,c,beta
In page “Profile” activate refinement of GW
In page “Corrections” activate refinement of “shift”;
Change number of Legendre terms for background to 30;
OK; Yes to rewrite
Press “Edit refinement commands”
Set 100 refinement cycles and Damping factor
OK
Press “Run Refine”
Refinement of profile slowly converges to the profile R value Rp~3.4% and GOF ~ 7.5.
Waiting for the full convergence is not necessary. Refinement can be interrupted by End or
Cancel - in this case End should be used in order to see the latest difference curve in the
powder profile viewer.
In page “Profile” activate refinement of GV
OK, YES to save changes
Run several cycles of Refine
Refinement of profile slowly converges to the profile R value Rp~3.0% and GOF ~ 7.0
For Gaussian parameters an important rule is that they cannot be refined all together
because they are linearly dependent. Usually we refine GW, then GV and finally GU. In our
case GU would not improve the fit.
“Show Powder profile”
The calculated profile fits quite well with the experimental one but some profiles are too
narrow, especially for low angles
Quit profile viewer
“Edit profile parameters”
In page “Profile” change the type of the peak-shape function to “Pseudo-Voigt” , activate
refinement of LX and LY
OK, YES to save changes
“Run Refine”
Refinement of profile slowly converges to the profile R value Rp~1.6% and GOF ~4.2.
“Show Powder profile”
Enlarge the profile and investigate details of the fit between the calculated and
experimental profile.
We are especially interesting whether asymmetry of peaks has been correctly described. Not
34 Jana2006 Cookbook, version October 2017
correctly described asymmetry can be found for low angle reflections:

“Edit profile parameters”

In page “Asymmetry/Diffractometer” select “correction by divergence”
Activate refinement of “HpS/L”, set HmS/L to zero and do not refine it
OK; YES to save changes

“Run Refine”
Refinement of profile slowly converges to the profile R value Rp=1.4 and GOF ~3.0.
“Show Powder profile”

Description of asymmetry seems to be better than before

Quit the profile viewer
Press "Save As" and make a backup copy of the structure
FINISH to close the Powder wizard

35 Jana2006 Cookbook, version October 2017

4. Space group determination
[On the screen: Tolerances for crystal system recognition]
Use default; NEXT
[On the screen: Select Laue symmetry]
Select the highest one – monoclinic; NEXT
[On the screen: Select cell centering]
Select P centering as it gives the best profile fit; NEXT
[On the screen: Select space group]
Select space group P21/a; NEXT
[On the screen: Final step of the space group test];
Accept the space group transformed into the original cell; FINISH
YES to refine profile parameters once more.
This new profile refinement uses only independent Bragg reflections compatible with P21/a.
The results are almost the same like for P-1 but calculation taking into the account the
symmetry is faster.

5. Structure solution
[On the screen: information that the solution wizard will start]
OK
Type formula Mg O10 H40 Br2 N8 C14; Formula units 2
Select “Use Superflip”
Check "Repeat Superflip: Number of runs" and choose 10 runs
With this setting Superflip will run ten times and a solution with the best fit will be used
OK
The agreement factors of the symmetry printed at the bottom of the Superflip listing will be
between 10-20, indicating rather worse fit for the screw axis.
Press the button “Draw structure”
The initial structure model should display octahedrally coordinated Mg located in one of
inversion centers, isolated Br, two oxygens of the water molecules, a fragment of the N4C7
cage and probably some false maxima. For instance:

Press button “Accept last solution”

Start File → Structure → Save as and make a backup copy

6. Completing structure without hydrogen atoms

36 Jana2006 Cookbook, version October 2017

In this part the task is to finalize the N4C7 cage from the difference Fourier.

Rietveld refinement with initial structure model could destroy profile parameters:
Start Parameters → Powder
Clear refinement keys for all refined powder parameters in page "Cell", "Profile",
"Asymmetry/Diffractometer" and "Corrections"

The stability of the structure model refined with powder data requires binding together ADP
parameters of some atoms and keeping them isotropic. For instance, carbon and nitrogen
atoms will all have the same isotropic ADP.
Plot the structure and note atoms around Mg
Start “Edit atoms”;
change the atoms around Mg to oxygen O1, O2, O3;
change all other atoms except Br and Mg to carbon.
Start Refinement options, page “Various”
Start “Restrictions” and define:
The same ADP for O*
The same ADP for C*

Be careful not to make by mistake identical their coordinates

OK;

In page Basic change number of refinement cycles to 100 and damping factor to 0.5
Clear “Disable atoms having too large isotropic ADP parameter”
Clear “Make only profile matching”
OK; Yes+Start
Rietveld refinement slowly converges with Robs=9%, Rp=6%, and GOF=13.
The numbers are only indicative!
In case of oscillations, note the label of the oscillating atom. Plot the structure and try to
identify and delete false maxima. In case of troubles, repeat charge flipping.
Change ADP of Br to anisotropic and repeat refinement
Br is slightly disordered and it can be described with an ellipsoid even from powder data
For isolated carbon atoms out of the fragment of the cage change their chemical type to
oxygen
Repeat refinement
Two of the isolated positions will be oxygens of water. The others will probably move to the
cage during the refinement.
Start Fourier option
In the page “Basic” define “F(obs)-F(calc) – difference Fourier)
OK; YES+START
OK to start procedure for including new atoms
The charge of difference maxima will be extremely low comparing with single crystal

37 Jana2006 Cookbook, version October 2017

diffraction, however some of them will be carbon or nitrogen atoms. It is advisable to plot
the maxima before refinement and to delete any strange position as false maxima.
Repeat refinement
Repeat plotting and deleting false maxima
Repeat difference Fourier
Repeat adding of atoms
Finally, the structure will be completed like this:

Rietveld refinement converges with Robs~3.6%, Rp~2.8%, and GOF~6.3.

ADP of carbon atoms of the cage may be negative at this stage
In the cage the atoms with three bonds are nitrogens.
Start “Edit atoms” and correct chemical types of the cage
Start Refinement options and add a restriction nitrogen atoms to have the same ADP:

Run refinement
Rietveld refinement converges with Robs~3.2%, Rp~2.4%, and GOF~5.2.
At this point ADP of nitrogens should be positive while ADP of carbons may still be close to
zero
Start “Edit atoms”
Add hydrogen atoms to carbons of the cage (automatic mode)
Run refinement
Rietveld refinement converges with Robs~3.2%, Rp~2.5%, and GOF~5.4.These value are
slightly worse than in the previous step but ADP of carbon atoms become positive
Start File → Structure → Save As and make a backup copy
Start Parameters → Powder
Activate refinement of all powder parameters which were fixed before starting with the
Rietveld refinement
Run refinement
Rietveld refinement converges with almost the same Robs, Rp and GOF.

38 Jana2006 Cookbook, version October 2017

All oxygen atoms belong to water. In order to include scattering power of hydrogen atoms
without adding them explicitly, their occupancy must be refined.
Activate refinement of occupancy of O1
Start refinement options, page “Various”
Start “Equations” and define the following equations:

Run refinement
Rietveld refinement converges with Robs~2.5%, Rp~2.0%, and GOF~4.0.These value are
significantly better.

Oxygen atoms of water, which is not coordinated to Mg, may have different ADP
Start refinement options, page Various
Start "Restrictions"
Delete the restriction for O*
Define separately the same ADP for O1, O2, O3 and for O4, O5:

Run refinement
The results are almost the same. Visual inspection of the powder profile shows good fit
between the calculated and observed lines.

39 Jana2006 Cookbook, version October 2017

40 Jana2006 Cookbook, version October 2017
Example 2.3.2: MagCag
Preparation of the rigid body template in M45
Revised: 4 March 2015
The purpose of this example is to prepare a template of the cage C7N4H14, which will be used
for the solution of the example 2.3.3 as a rigid body. The template will be defined in the file
M45 and used in solution of the example 2.3.3.

The preparation of M45 has no special support in the current version of Jana2006. Part of the
job must be done with a text editor.

In the following steps we shall use the CSD structure HICTOD and we shall prepare two versions
of the cage: with C1 symmetry and Cs symmetry.

Input files: HICTOD.cif from the CSD database

Expected chemical scheme:

1. Reading structure from CIF

Start Jana2006
“File → Structure → New” opens a file manager
Left pane: locate directory the input file
Right pane: double-click HICTOD
[On the screen: Specify type of the file to be imported]
Select : Structure → From CIF; NEXT
Select: HICTOD.CIF
NO to import data

2. Deleting symmetry equivalent atoms

It happens with some CSD CIF files that symmetry equivalent atoms are also present, which
leads to their multiple occurrences when symmetry is applied. In the next step, we shall
delete these atoms
Start "Edit Atoms"
Select atoms ending with “g” and “d”
Start "Action → Delete selected atoms"
OK; YES to save changes

41 Jana2006 Cookbook, version October 2017

3. Getting the cage from HICTOD
Start "Plot structure"; "Draw + continue"
Use , and to plot atoms from M40
Use Picture → Viewing direction → c
Use to display the unit cell
Use to complete the fragments
Now the cage is clearly visible
The atoms of the cage are symmetry related by a mirror plane at y=0.25
Press Undo, to see again only the symmetry independent atoms
Select atoms (symmetry independent) not belonging the cage
Using the right mouse button activate ploting of their labels; note the labels
Quit Diamond and return to Jana
Start "Edit Atoms"
Delete the atoms not belonging to the cage; OK; YES to rewrite changes
Start "Plot structure" and verify that only atoms of the cage are present
Return to Jana

4. Transformations
M45 templates contain information about the point group symmetry. The axis c is used as
the unique axis of the point group. For this reason the unit cell taken from CSD must be
transformed in order to get the mirror plane of the cage perpendicular to c and located at
z=0. Moreover, all atoms of the cage (not only symmetry independent atoms) must be
present in the template.
Start "Tools → Transformations → Origin shift"
Move the origin to point 0, 0.25, 0
Now the mirror plane of the cage is located at y=0 (can be verified in M40)
Start "Edit Atoms"
Press "Select all"
Start "Action → Expand selected atoms"
Press "Next matrix" ; "Fill by a symmetry operation"
Select "x, -y, z"; OK;
OK; OK; YES to rewrite changes
Start "EditM50"
Change symmetry to P1; OK; YES to rewrite M50
Start "Edit Atoms"
Press "Select all"
Start "Action → Merge selected atoms"; use default merging distance;
The program merges 7 atoms, resulting in 25 atoms of the cage
Start "Tools → Transformations → Cell transformation"
Press "by a matrix"
Define transformation matrix -1 0 0 / 0 0 1 / 0 1 0
OK; OK; select "rewrite the old structure"; OK
Start "Edit Atoms"
Press "Select all"
Start "Action → Transform selected atoms"
Use unit transformation matrix and translation vector 1, 0, 0
This transforms the x coordinates back to the unit cell
42 Jana2006 Cookbook, version October 2017
 Start "Plot structure"; "Draw + continue"
Use , and to plot atoms from M40
Use Picture → Viewing direction → b
Use and to display the unit cell
This shows the resulting cage of 25 atoms with mirror plane located at z=0

5. Labels
Quit Diamond
Start "Edit Atoms"
Double-click the first atom
Using "Define" and "Edit" page change the suffix to "m" if z=0, "1" if z<0 and "2" if z>0
Arrows close to the "List" button can be used to move between atoms in the editing mode
Renaming can also be done in a text editor (Edit/View → Edit of M40 file).
The resulting labels will have "m" for atoms at the mirror plane, "1" for atoms below the
mirror plane and "2" for atoms above the mirror plane. The suffixes "1" and "2" resulted
from the previous expansion by the mirror plane.

43 Jana2006 Cookbook, version October 2017

6. Text editor
Quit Diamond and Jana
Copy M40 to Magcag-symmC1.m45
Open Magcag-symmC1.m45 in a text editor
Delete five header lines
Delete the s.u. block
Delete every second line (i.e. the lines without atom label)
Add two spaces and "0" at the end of each line
Add the following header lines:
"cell 7.585 10.837 16.065 90 90 90" (this can be copied from M50)
"pgroup C1"
"scdist C-C 1.52"
Save the file and quit the editor.

This is the resulting file Magcag-symmC1.m45:

cell 7.585 10.837 16.065 90 90 90

pgroup C1
scdist C-C 1.52
N1_m 3 1 1.000000 0.250500 0.132420 0.000000 0
N2_m 3 1 1.000000 0.194100-0.129200 0.000000 0
N3_1 3 1 1.000000 0.011590 0.024260-0.075730 0
N3_2 3 1 1.000000 0.011590 0.024260 0.075730 0
C1_m 1 1 1.000000 0.373700-0.084900 0.000000 0
C2_m 1 1 1.000000 0.403100 0.054200 0.000000 0
C3_1 1 1 1.000000 0.142500 0.125210-0.074590 0
C3_2 1 1 1.000000 0.142500 0.125210 0.074590 0
H1_1 2 1 1.000000 0.219800 0.115500-0.122300 0
H1_2 2 1 1.000000 0.219800 0.115500 0.122300 0
H2_1 2 1 1.000000 0.080600 0.202900-0.081200 0
H2_2 2 1 1.000000 0.080600 0.202900 0.081200 0
C4_m 1 1 1.000000-0.095500 0.035800 0.000000 0
H3_m 2 1 1.000000-0.185600-0.027800 0.000000 0
H4_m 2 1 1.000000-0.154200 0.115300 0.000000 0
C5_1 1 1 1.000000 0.093300-0.099710-0.074750 0
C5_2 1 1 1.000000 0.093300-0.099710 0.074750 0
H5_1 2 1 1.000000 0.000900-0.160900-0.081200 0
H5_2 2 1 1.000000 0.000900-0.160900 0.081200 0
H6_1 2 1 1.000000 0.170800-0.106700-0.122500 0
H6_2 2 1 1.000000 0.170800-0.106700 0.122500 0
H9_2 2 1 1.000000 0.427000-0.121100 0.053300 0
H9_1 2 1 1.000000 0.427000-0.121100-0.053300 0
H10_2 2 1 1.000000 0.473000 0.074000 0.051500 0
H10_1 2 1 1.000000 0.473000 0.074000-0.051500 0

The keyword scdist is used for scaling of the complete object using arbitrary number
described with arbitrary identifier (without spaces). In our case the identifier is “C-C” and the
number is a typical C-C distance in the cage. But the identifier could be also “scale” and the
number could be 100.
Each line except header lines contains the following items: label, number of chemical type
(indicates that chemical types are different - they will be specified during application of the
rigid body in Jana2006), unused, site occupancy, x, y, z, symmetry identifier (0 means
symmetry independent with respect to the used point group).

44 Jana2006 Cookbook, version October 2017

Copy Magcag-symmC1.m45 to Magcag-symmCs.m45
Open Magcag-symmCs.m45 in a text editor
Replace the point group C1 with Cs
For atoms with the suffix "1", i.e. those with z<0, replace "0" at the end of each line with
"1". This indicates the atoms are symmetry dependent
For atoms with the suffix "m" change the site occupancy to 0.5, because they are located at
mirror plane

This is the resulting file Magcag-symmCs.m45:

cell 7.585 10.837 16.065 90 90 90

pgroup Cs
scdist C-C 1.52
N1_m 3 1 0.500000 0.250500 0.132420 0.000000 0
N2_m 3 1 0.500000 0.194100-0.129200 0.000000 0
N3_1 3 1 1.000000 0.011590 0.024260-0.075730 1
N3_2 3 1 1.000000 0.011590 0.024260 0.075730 0
C1_m 1 1 0.500000 0.373700-0.084900 0.000000 0
C2_m 1 1 0.500000 0.403100 0.054200 0.000000 0
C3_1 1 1 1.000000 0.142500 0.125210-0.074590 1
C3_2 1 1 1.000000 0.142500 0.125210 0.074590 0
H1_1 2 1 1.000000 0.219800 0.115500-0.122300 1
H1_2 2 1 1.000000 0.219800 0.115500 0.122300 0
H2_1 2 1 1.000000 0.080600 0.202900-0.081200 1
H2_2 2 1 1.000000 0.080600 0.202900 0.081200 0
C4_m 1 1 0.500000-0.095500 0.035800 0.000000 0
H3_m 2 1 0.500000-0.185600-0.027800 0.000000 0
H4_m 2 1 0.500000-0.154200 0.115300 0.000000 0
C5_1 1 1 1.000000 0.093300-0.099710-0.074750 1
C5_2 1 1 1.000000 0.093300-0.099710 0.074750 0
H5_1 2 1 1.000000 0.000900-0.160900-0.081200 1
H5_2 2 1 1.000000 0.000900-0.160900 0.081200 0
H6_1 2 1 1.000000 0.170800-0.106700-0.122500 1
H6_2 2 1 1.000000 0.170800-0.106700 0.122500 0
H9_2 2 1 1.000000 0.427000-0.121100 0.053300 0
H9_1 2 1 1.000000 0.427000-0.121100-0.053300 1
H10_2 2 1 1.000000 0.473000 0.074000 0.051500 0
H10_1 2 1 1.000000 0.473000 0.074000-0.051500 1

The created templates, Magcag-symmC1.m45 and Magcag-symmCs.m45, could be

accompanied with plots in jpeg format, Magcag-symmC1.jpg and Magcag-symmCs.jpg. In
our example we don’t need them.

45 Jana2006 Cookbook, version October 2017

46 Jana2006 Cookbook, version October 2017
Example 2.3.3: MagCag
Solution of simple organometallic structure from powder data, using the rigid
body template from M45
Revised: 4 October 2015
Powder data measured with diffractometer Empyrean (PANalytical) using CuKα radiation,
focusing mirror and Debye-Scherrer geometry (sample in capillary), at ambient temperature.

Input files: magcag.xrdml with powder profile

Magcag-symmC1.m45 and Magcag-symmCs.m45: rigid body templates prepared in Example
2.3.2
Additional information:
unit cell 9.19 21.38 8.12 90 113.78 90
Expected scheme:

1. Creating new jobname

Start Jana2006
“File → Structure → New” opens a file manager
Left pane: locate directory with input files
Right pane: double-click magcag

2. Import procedure
Select “Powder data – various CW formats”; NEXT
[On the screen: Powder data form]
Select "PANalytical XRDML"
For method select “Debye-Scherrer”
[On the screen: Complete/correct experimental parameters]
Fill cell parameters
Check the radiation (should be interpreted from the xrdml file as Copper with doublet):

47 Jana2006 Cookbook, version October 2017

 NEXT; FINISH; OK to close data repository; YES to accept the data set
[On the screen: question how to continue]
Select “Yes, I would like to continue with the wizard”; OK
The typical steps for refinement of powder profile will be offered in the wizard. However, all
steps can be also done separately from the basic window of Jana2006.

3. Refinement of the powder profile

[On the screen: refinement of the powder profile by the le Bail algorithm]
Press “Show powder profile”
Press “Default” to see more details
Press “Create new” to create a manual background
[On the screen: Options for generating background profile]
For “Number of manual background points” set 60
OK; Quit powder viewer
[On the screen: A manual background has been created]
Select “Accept the new background” and “Reset polynomial coefficients to zero”
The manual background was created because with this data background calculated with
Legendre polynomials would not be correct for high angle reflections.(For detail see example
2.4 – PFPhenyl)
Press “Edit profile parameters”
In page “Profile” activate refinement of GW
Unlike for structure parameters, refinement keys of profile parameters are not set
automatically
In page Corrections:
Set 30 for “Number of terms” of Legendre polynomials
Polynomial background will be refined in addition to the manual one
Activate refinement of “shift”;
OK; Yes to rewrite
Press “Edit refinement commands”
Set 100 refinement cycles
OK
Press “Run Refine”
Refinement of profile slowly converges to the profile R value Rp~2.9% and GOF ~ 7.3.
Waiting for the full convergence is not necessary. Refinement can be interrupted by End or
Cancel - in this case End should be used in order to see the latest difference curve in the
powder profile viewer.
“Show Powder profile”; “Default”

48 Jana2006 Cookbook, version October 2017

Move the slider towards low angles; “Fit Y”
The calculated profile fits quite well with the experimental one but some profiles are too
narrow, especially for low angles
Quit profile viewer
“Edit profile parameters”
In page “Cell” activate refinement of a,b,c,beta
In page “Profile” activate refinement of GV
OK, YES to save changes
Run several cycles of Refine
Refinement of profile slowly converges to the profile R value Rp~2.8% and GOF ~ 7
For Gaussian parameters an important rule is that they cannot be refined all together
because they are linearly dependent. Usually we refine GW, then GV and finally GU. In our
case GU would not improve the fit.
In page “Profile” change the type of the peak-shape function to “Pseudo-Voigt” , activate
refinement of LX and LY
OK, YES to save changes
“Run Refine”
In case of oscillations press “Parameters” (during the refinement) and change “Damping
factor” to 0.5
Refinement of profile slowly converges to the profile R value Rp~1.7% and GOF ~4.3.
“Show Powder profile”
Enlarge the profile and investigate details of the fit between the calculated and
experimental profile.
We are especially interesting whether asymmetry of peaks has been correctly described. Not
correctly described asymmetry can be found for low angle reflections:

“Edit profile parameters”

In page “Asymmetry/Diffractometer” select “correction by divergence”
Activate refinement of “HpS/L” (do not refine HmS/L)
“Run Refine”
Refinement of profile slowly converges to the profile R value Rp=1.4% and GOF ~3.1. “Show
Powder profile”

49 Jana2006 Cookbook, version October 2017

 The description of asymmetry for the low angle peak is considerably improved:

Quit the profile viewer

Press "Save As" and make a backup copy of the structure
FINISH to close the Powder wizard

4. Space group determination

[On the screen: Tolerances for crystal system recognition]
Use default; NEXT
[On the screen: Select Laue symmetry]
Select the highest one – monoclinic; NEXT
[On the screen: Select cell centering]
Select P centering as it gives the best profile fit; NEXT
[On the screen: Select space group]
Select space group P21/a; NEXT
[On the screen: Final step of the space group test];
Accept the space group transformed into the original cell; FINISH
YES to refine profile parameters once more.
This new profile refinement uses only independent Bragg reflections compatible with P21/a.
The results are almost the same like for P-1 but calculation taking into the account the
symmetry is faster.

5. Structure solution
[On the screen: information that the solution wizard will start]
OK
Type formula Mg O10 H40 Br2 N8 C14; Formula units 2
Select “Use Superflip”
Check "Repeat Superflip: Number of runs" and choose 10 runs
With this setting Superflip will run ten times and a solution with best R value will be used
“Run Solution”
In case that Superflip does not converge interrupt the calculation, discard the results and
repeat it. It will start from a new set of random phases and the result may be different. The
agreement factors of the symmetry printed at the bottom of the Superflip listing will be
between 10-20, indicating rather worse fit for the screw axis.

50 Jana2006 Cookbook, version October 2017

 Press the button “Draw structure”
The initial structure model contains octahedrally coordinated Mg located in one of inversion
centers, isolated Br, two oxygens of the water molecules, a fragment of the N4C7 cage and
probably some false maxima. For instance:

Press button “Accept last solution”

6. Preparation of the starting Rietveld refinement

Start “Edit atoms” and change chemical types of atoms coordinated to Mg to oxygen
Change to oxygen also the two carbon atoms, which are in fact lattice water molecules
Rename the changed atoms according to their chemical types
OK; YES to save changes
Start “Refine commads”; go to page “Basic”
Set 100 refinement cycles with Damping factor 0.1
Uncheck “Make only profile matching”
OK; YES to save changes without starting refinement
Start “Edit profile”
Fix all profile parameters: a, b, c, beta, GW, GP, LX, LY, HpS/L, shift
This means that only Legendre background will be refined
OK, YES to save changes
Start File → Structure → Save as and make a backup copy named “start”

7. Completing the cage

Start “Refine commands”; go to page “Various”
Press “Fixed commands”
Fix all parameters for O* Mg* Br*; OK
Press “Restrictions”
Restrict ADP of C* to be the same; OK
Be careful not to restrict coordinates!
OK; YES+Start
During refinement you may change damping from 0.1 to 0.5
Refinement converges with R~11%, Rp~10%, GOF~23
Run difference Fourier, YES to start procedure for inserting atoms
Insert new maxima in reasonable distances to the atoms of the cage
Repeat refinement
Repeat difference Fourier and inserting of maxima until 11 carbon atoms is present

51 Jana2006 Cookbook, version October 2017

 “Reasonable distance” should not be taken too strictly for Fourier maxima from powder
data
Refinement converges with R~5%, Rp~4%, GOF~8
Start File → Structure → Save as and make a backup copy named “cage_atomic”

8. Placing the cage template with point group symmetry C1

Start “Plot structure”
Using the button plot the atoms of the cage and ensure they make a symmetry
contiguous motif. If not, return to Jana, start Edit atoms and put C* to symmetry contiguous
motif. Otherwise placing of the molecule will not work.

Plot the cage and compare with the template prepared in the previous example. Probably
C9 corresponds to C1_m; C3 corresponds to C2_m; C8 corresponds to C4_m (your labels
can be different!)
Return to Jana2006
Start “Parameters → Molecules → New molecule”
[On the screen: Atoms of the new molecule from:”]
Select “Model file”
For name of the molecule type “kage”
We need a name not starting with C, N, O, B, H and M
For “Model name” browse for “Magcag-symmC1.m45” created in the previous example
Keep the scaling distance unchanged
“Show the model molecule” is disabled because we don’t have Magcag-symmC1.jpg
NEXT
[On the screen: Specify atomic types in the model molecule]
Do not change the suggested atomic types; NEXT
[On the screen: Define the molecular reference point]
Select “Explicit”
In “Reference point” textbox type N1_m
The reference point should be some atom from the mirror plane in order to reflect the
symmetry of the molecule. Such atoms in the model molecule have suffix _m (see Example
2.3.2)
NEXT
[On the screen: Define and complete molecular position #1]
For the pairs “model atoms” – “actual positions” choose c1_m and C9; C2_m and C3; C4_m
and C8 (with actual names taken previously from Diamond)

52 Jana2006 Cookbook, version October 2017

 For “Maximal coincidence distance” type 0.5
The coincidence ration should be 11/25 but it depends on the previous refinement
Try to adjust the coincidence distance to reach 11 coinciding carbons.
If carbons of the model molecule coincide e.g. with oxygens this means that the cage has
been positioned incorrectly, probably because the carbon atoms in the atomic part do not
form a symmetry contiguous motif.
Press “Apply+End”
Start “Edit atoms”
Ensure that the atomic part does not contain carbon atoms.
OK, YES to save changes
Start “Plot structure”
The structure with the cage should look like this (depending which carbon positions were
used for placing the cage):

In this example one water oxygen is too close to the cage but it may change during the
refinement

Return to Jana2006
Start “Refine commands”; go to page “Various”
Press “Fixed commands”
Fix “all parameters” for C*, H*, N* (i.e. for atoms of the cage)
Because the molecule is called “kage”, the wild character C* does not include the molecule
and therefore molecular parameters (rotations and translation of the cage), will not be
fixed.
OK; OK; YES+START to save changes and start refinement
During refinement you may change damping from 0.1 to 0.5
After many cycles refinement converges with R~4.5%, Rp~4%, GOF~9
Start “Plot structure”
The cage should be now in better position, i.e. without short distances to lattice water

9. Refinement of Mg, O and Br

Start “Refine commands”; go to page “Various”
Press “Fixed commands”
Disable or delete the “fixed all” commands for Mg* O* Br*
OK
53 Jana2006 Cookbook, version October 2017
 Go to page “Equations” and define the following equations:

All oxygens belong to water; refinement of their occupancy mimics the missing hydrogens
OK
Go to page “Various”, press “restrictions”
Define the same ADP for all oxygen atoms (O*); OK
Be careful not to make identical coordinates!
OK; YES to save commands without starting Refine
Start “Edit atoms”
Set harmonic ADP for Br1
This works because Br is slightly disordered
Activate refinement of occupancy (ai) of O1
OK, YES to save changes
Start “Refine”
The refinement converges with R~3%, Rp~2.3%, GOF~5

10. Refinement of atoms of the cage

Start “Edit atoms”
Select all carbon atoms of the cage (blue carbons)
Press “Action → Adding of hydrogen atoms”
This procedure will not add hydrogens, because they are already present, but it will
introduce Keep commands necessary for fixing of their geometry
OK; YES to save changes
Start “Refine commads”; go to page “Various”
Press “Restrictions”
Define the same ADP for all nitrogen atoms (N*)
Be careful not to make identical coordinates!
OK
Press “Fixed commands”
Disable or delete the “fixed all” command for C* N* H*
OK; go to page Basic
Set “Damping factor” 0.1
OK; YES+START
The refinement converges with R~2.8%, Rp~2.1%, GOF~4.3
Start “Plot structure” verify the structure

Final plot:

54 Jana2006 Cookbook, version October 2017

 Make backup copy

11. Placing the cage template with point group symmetry Cs

Now we shall test how the cage with Cs point symmetry would fit
Start "Edit atoms"
Select the model molecule (blue atoms)
Start "Action → Atoms from molecule to atomic part"
OK; YES to rewrite changes
The molecule had only one actual position and therefore this atomic structure model is
equivalent with the previous one
Start "Refine"
R values must be the same like with the molecule
Now we shall place the template Cs in such a way that it coincides with the cage present in
the atomic part
This is the template how it was created in the example 2.3.2:

Start "Plot structure" and try to orient your cage similarly. Make use of the fact that there is
only one C-C bond in the cage.

55 Jana2006 Cookbook, version October 2017

 From the comparison we can see which pairs of atoms should be used for placing the
molecule
Return to Jana2006
Start "Parameters → Molecule → New molecule"
[On the screen: Atoms of the new molecule from:”]
Select “Model file”
For name of the molecule type “kage”
For “Model name” browse for “Magcag-symmCs.m45” created in the previous example
Keep the scaling distance unchanged
“Show the model molecule” is disabled because we don’t have Magcag-symmC1.jpg
NEXT
[On the screen: Specify atomic types in the model molecule]
Do not change the defaults atomic types; NEXT
[On the screen: Define the molecular reference point]
Select “Explicit”
In “Reference point” textbox type N1_m
This should be some atom from the mirror plane
NEXT
[On the screen: Define and complete molecular position #1]
For the pairs “model atoms” – “actual positions” type the three previously noted pairs
For "Maximal coincidence distance" type 0.6
Press "Show coinciding atoms"
We should see the complete coincidence 25/25 and the coinciding atoms should have the
same chemical types.

Press “Apply+End”
Start "Edit atoms"
The atomic part should contain only Br1, Mg1 and five oxygens
Start "Plot structure"
The cage should be correctly placed – verify by plotting
In a plot slash "/" means the atoms is transformed by the point group symmetry of the
molecule, while "a" indicates the first actual position of the molecule.

12. Refinement of the cage with Cs point symmetry

Make backup copy
Start "Refine commands"; go to page "Various"
Press "Keep commands" and delete all of them

56 Jana2006 Cookbook, version October 2017

OK; OK; YES to save commands without starting Refine
Start "Edit Atoms"
Select all carbons
Start "Action → Adding of hydrogen atoms"
Select "Try automatic run"
The procedure should not add any new hydrogen atoms
OK; YES to quit “Edit atoms”and save changes
Start "Refine Commands"
Set zero refinement cycles; OK; YES+START
This applies the previously created keep commands
Start "Plot structure" and verify hydrogen positions
Start "Refine commands", set 100 refinement cycles with Damping factor 0.5
OK; YES+START to save commands and run refinement
The refinement has difficulties to reach convergence
The fit is R~3.2%, Rp~2.3%, GOF~5, 53 structure parameters; i.e. only slightly worse fit
comparing with the asymmetric cage

57 Jana2006 Cookbook, version October 2017

58 Jana2006 Cookbook, version October 2017
Example 2.4: PFPhenyl
Refinement of moderately sized organometallic structure from powder data
with help of geometry constrains.
Revised: 20 April 2017
Powder data measured with diffractometer Empyrean (PANalytical) using CuKα1 radiation and
Debye-Scherrer geometry (sample in capillary), at ambient temperature.

Input files:
pfphenyl.xrdml with powder profile
pfphenyl.cif with solution obtained by the program FOX
Additional information:
unit cell 10.77024 9.957075 26.70771 90 91.683 90
expected symmetry P21/n
composition C30 N3 H29 P O2 F3 (Z=4)
Chemical scheme:

1. Creating new jobname

Start Jana2006
“File → Structure → New” opens a file manager
Left pane: locate directory with input files
Right textbox: type "pfphenyl"; OK

2. Import procedure
[On the screen: Specify type of the file to be imported]
Select “Structure – from CIF”; NEXT
[On the screen: select input CIF file]
Select pfphenyl.cif; OK
[On the screen: CIF does not contain any reflection block]
YES to import data from file
[On the screen: Specify type of the file to be imported]
Select "various CW formats"; NEXT
[On the screen: Powder data from:]
Select "PANalytical XRDML"
For method select “Debye-Scherrer”; NEXT
[On the screen: Complete/correct experimental parameters]

59 Jana2006 Cookbook, version October 2017

 Cell parameters were taken from CIF
Radiation was interpreted from the xrdml file as CuKα1

NEXT; FINISH; OK to close data repository; YES to accept the data set

3. Refinement of the powder profile

For refinement of the powder profile we will use the le Bail wizard.
Start "Wizards → le Bail wizard"
[On the screen: Refinement of the powder profile by the le Bail algorithm]
Press “Show powder profile” – just to see how it looks like
Press “Default” to see more details
Quit powder viewer
Press “Edit profile parameters”
In page “Profile” activate refinement of GW
Unlike for structure parameters, refinement keys of profile parameters are not set
automatically
In page Corrections:
Activate refinement of “shift”;
Change number of Legendre terms for background to 30;
OK; Yes to rewrite
Press “Edit refinement commands”
Make sure that "Make only profile matching" is activated
This means that the structure imported from CIF is not used in the refinement
Set 100 refinement cycles
OK;
Press “Run Refine”
Refinement of profile slowly converges to the profile R value Rp~2.9% and GOF ~ 3.5.
Waiting for the full convergence is not necessary. Refinement can be interrupted by End or
Cancel - in this case End should be used in order to see the latest difference curve in the
powder profile viewer.
“Show Powder profile”
Press "Default"
Move the theta slider towards high angles
60 Jana2006 Cookbook, version October 2017
Press "Details ON"; YES to plot many component functions
Press "Fit Y"
The calculated profiles show problem with background. The background is too low and the
Bragg peaks are too high, giving together very nice fit between calculated and observed line,
which is unfortunately false. We shall therefore use the manual background.

Press "Details OFF"

Press "Create new" button (lower right corner of the Powder profile viewer)
[On the screen: Options for generating background profile]
Use 60 background points and significance intensity level 1; OK
Using the slider check the manual background. It can be edited by dragging the points by
mouse.
Press "Quit"
[On the screen: A manual background has been created]
Select "Accept the new manual background" and "Suppress using of polynomial terms"
OK
Now we are back in the le Bail wizard
Press "Run Refine"
The refinement swiftly converges with Rp~2.8%, GOF~3.6.
“Edit profile parameters”
In page “Cell” activate refinement of a,b,c,beta
In page “Profile” activate refinement of GV
OK, YES to save changes
Run Refine
Refinement of profile converges to the profile R value Rp~2.5% and GOF ~ 3.1
For Gaussian parameters an important rule is that they cannot be refined all together
because they are linearly dependent. Usually we refine GW, then GV and finally GU. In our
case GU would not improve the fit.
“Edit profile parameters”
In page “Profile” change the type of the peak-shape function to “Pseudo-Voigt” , activate
refinement of LY
OK, YES to save changes
“Run Refine”
Refinement of profile slowly converges to the profile R value Rp~2.5% and GOF ~3.0.
“Show Powder profile”
Enlarge the profile and investigate details of the fit between the calculated and
experimental profile.

61 Jana2006 Cookbook, version October 2017

 Some low angle peaks exhibit strong asymmetry:

“Edit profile parameters”

In page “Asymmetry/Diffractometer” select “correction by divergence”
Activate refinement of “HpS/L” and fix to zero “HmS/L”
OK; YES to save changes
"Edit refinement commands"
In page "Basic" set "Damping factor" to 0.5
OK; “Run Refine”
Refinement of profile slowly converges to the profile R value Rp=1.9% and GOF ~1.8.
Waiting for full convergence is not necessary.
“Show Powder profile”
Asymmetry seems to be well described
Quit the profile viewer
Press "Save As" and make a backup copy of the structure
FINISH to close the Powder wizard

4. Verification of symmetry
Start File → Reflection file → Make space group test
YES to continue
[On the screen: Tolerances for crystal system recognition]
Use default; NEXT
[On the screen: Select Laue symmetry]
Select the highest one – monoclinic; NEXT
[On the screen: Select cell centering]
(calculation takes time! It is finished when the program prints warning about cell centering
at the end of the list of centering symbols).
The table shows Rp values calculated for various centering vectors. For instance, for A
centering the program discards corresponding extinct Bragg reflections and calculates Rp. If
A centering would really exist, the new Rp would be similar like the one for primitive unit cell
because the discarded forbidden reflections would not be vital for the profile fit. In our case,
A centering is not possible because the corresponding Rp is much worse than for the
primitive unit cell.
Select P centering as it gives the best profile fit; NEXT
[On the screen: Information about progress of space group tests. It takes some time…]
[On the screen: Select space group]

62 Jana2006 Cookbook, version October 2017

 Among the space groups with good profile fit we are looking – in analogy with a single-
crystal experiment – for a space group with maximum number of extinct reflections which
are not needed for profile fit. The arguments are the same like for selecting the centering.

The space group P21/n, which is the space group used for the FOX solution has good FOM
(figure of merit). This space group has been already entered with the CIF structure so we can
cancel the process at this point. We could also consider P2/n but we will rely on the FOX
symmetry.
CANCEL

5. Refinement of the free atomic model

Make sure that the backup copy really exists
Start "Plot structure"
In Diamond use "Build → Connectivity" and suppress F-H and F-O bonds
Use to plot the structure
The structure from FOX solution is looking like this:

Before refinement we have to fix geometry for hydrogen atoms. Hydrogen positions cannot
be determined from this powder data set.

Return to Jana
Start "Edit atoms"
Select all carbon atoms and N9
Start "Action → Adding of hydrogen atoms"
[On the screen: Adding of "hydrogen" atoms"]
Use default settings; OK
Hydrogen atoms of carbon have been fixed automatically while for N9 the program offers a
dialogue
[On the screen: Adding "hydrogen" atoms for "N9"]
Select "Trigonal"; C28 and P1 for neighbors; H54 for hydrogen
Press "Apply → Go to next"
OK; YES to save changes
Start Refinement options (Right-click on the Refine icon)
Select the page "Basic"
Set 0 refinement cycles
Uncheck "Make only profile matching"
This changes le Bail fitting into Rietveld refinement, where intensities of the profile points
are calculated from the structure model

63 Jana2006 Cookbook, version October 2017

 OK; YES+START
R values are bad because no scales were refined. However, positions of hydrogen atoms
were updated according to the previously introduced restrains and now we can check in the
plotting program correctness of hydrogen positions.
Start "Plot structure"
Verify hydrogen positions
Quit Diamond
Start "Edit Profile" and fix all profile parameters: a,b,c,beta,GV,GW,LY,HpS/L and shift.
OK, YES to save changes
Using "File → Structure → Save as" make a backup copy named "pfphenyl_free"
Start Refinement options; set 100 refinement cycles; set Damping factor 0.2
In page "Various" start "Restrictions"
Set the same ADP for all carbon atoms (C*)
Set the same ADP for all nitrogen atoms (N*)
Set the same ADP for all oxygen atoms (O*)
Set the same ADP for all fluorine atoms (F*)

OK; YES (i.e. DO NOT start refinement)

Using "File → Structure → Save as" make a backup copy
Start "Refine"
Now the structure is refined but the originally obtained profile parameters are fixed
The refinement converges slowly to Rp~2.2%, Robs~6.5% and GOF~2.
Start "Plot structure" and plot the refined structure

We can see that the PN3O group has been refined correctly but phenyl groups are not
planar and the CF3 group has been destroyed
6. Fixing geometry with keep commands
Start "Tools → Recover files"
This makes one step back. Otherwise "File → Structure → Copy in" would be used for
importing the previously created backup file.
Start "Plot structure" in order to see atom labels.
With "Draw+continue" Jana2006 and Diamond can be used together and atom labels seen

64 Jana2006 Cookbook, version October 2017

in Diamond can be immediately applied in Jana dialogues, provided that the computer
screen is large enough. Otherwise please note the labels for later use or use the figure from
the Cookbook

Start Refine Commands

In page Various start "Keep commands"
The already existing 50 commands keep the geometry of hydrogen atoms
Select "Geometry" and "Plane"
For "List of atoms" type labels of carbons in one of the Phenyl rings (without hydrogens)
C17 C20 C23 C27 C31 C25
Press "Add"
Repeat the process for the remaining Phenyl rings
C14 C18 C30 C32 C26 C21
C57 C56 C58 C60 C61 C59
C10 C11 C12 C29 C22 C16
Switch from "Plane" to "Rigid"
For "List of atoms" type C28 O6 C24 F2 F4 F5
Press "Add"
The introduced keep commands:

OK; OK; YES+START

The refinement converges slowly to Rp~2.6%, R~7.6% and GOF~2.4, i.e. slightly worse results

65 Jana2006 Cookbook, version October 2017

 comparing with unrestricted geometry
Damping factor may be changed to 1
Start "Plot structure" and verify the result
The geometry is now much more reasonable, but some angles in the phenyl rings are too
different of 120º.
Make backup copy of this refinement named (for instance) phphenyl_keepgeom

7. Application of rigid body to phenyl rings

Start "Refinement options", go to the page "Various"
Start "Keep commands" and delete or disable the four "keep geom plane" commands ("keep
geom rigid" should be retained)
OK; OK; YES to save commands without starting Refine
Start "Parameters → Molecules → New molecule"
[On the screen: Atoms of the new molecule from:]
Select "Atomic part"
Write "Phenyl" for the name of the molecule
NEXT
[On the screen: Select atoms for the molecule]
Select C10, C11, C12, C29, C22, C16, H46, H50, H33, H34, H35 (i.e. one of the phenyl rings)
OK
[On the screen: Define the molecular reference point:]
Select Geom. center
For molecules not located in a special position the position of the reference point is not
important. For symmetrical molecules the reference point is used for derivation of symmetry
restrictions
NEXT
[On the screen: Complete molecular position #1:]
Use Occupancy 1
The Coincidence ratio 11/11 means that all atoms of the actual position #1 coincide with
atoms in the atomic part. This is obviously fulfilled, because the model molecule has been
taken from the atomic part
Press "Apply+Next position"

66 Jana2006 Cookbook, version October 2017

[On the screen: Define and complete the molecular position #2:]
For model atoms select three carbons from the model molecule: C16, C11 and C29
For the actual position choose three corresponding carbons from another existing phenyl
group, for instance C57, C58 and C61
For "Maximal coincidence distance" type 0.6 (in Angstroms)
Press "Show coinciding atoms"
It shows that the actual molecular position #2 coincides with another 11 atoms from the
atomic part. These atoms will be automatically deleted in the atomic part.

Press "Apply+Next position"

For model atoms re-use C16, C11 and C29
For actual position use C26, C30 and C14
Number of coinciding atoms should be again 11
Press "Apply+Next position"
For model atoms re-use C16, C11 and C29
For the last actual position use C20, C25 and C27
Number of coinciding atoms should be again 11
Press "Apply+End"
Start "Edit atoms"
The free atoms are black, the atoms of the model molecule are blue
OK
Start "Plot structure"
The structure should look more or less like before (with the difference that now we have four
identical phenyl groups)

The old keep commands acting for atoms replaced by the actual molecular positions, except
of those selected as a model molecule, were automatically deleted. The atoms will be still
kept in ideal position as before but now through the keep commands applied to the model
molecule.

However, we have also used keep commands involving atoms from the phenyl groups and
N-CH2- group. The command defining positions of two hydrogen atoms bonded to C19 has
not been deleted, and it still contains the atom C16, which belongs to the model molecule
and not to some actual position. Therefore, C16 must be changed to C16a:

67 Jana2006 Cookbook, version October 2017

 Start Refine Commands
In page Various start "Keep commands" and load the relevant keep command:

Change C16 to C16a; Add

OK; OK; YES to save commands without starting Refine

The commands for keeping the geometry of hydrogen atoms in N-CH2- groups for remaining
three atoms (C55, C15 and C13) have to be re-defined:

Start "Edit atoms"

Select C55, C15 and C13
Start "Action → Adding of hydrogen atoms"
Use default settings; OK
(number of atoms should not change)

Here are the final keep commands:

keep hydro triang C10 2 1 C16 C11 0.96 H33
keep ADP riding C10 1.2 H33
keep hydro triang C11 2 1 C12 C10 0.96 H34
keep ADP riding C11 1.2 H34
keep hydro triang C12 2 1 C11 C29 0.96 H35
keep ADP riding C12 1.2 H35
keep ADP riding C13 1.2 H36 H37
keep ADP riding C15 1.2 H40 H39
keep hydro tetrahed C19 2 2 N7 C16a 0.96 H44 H43
keep ADP riding C19 1.2 H44 H43
keep hydro triang C22 2 1 C29 C16 0.96 H46
keep ADP riding C22 1.2 H46
keep hydro triang C29 2 1 C22 C12 0.96 H50
keep ADP riding C29 1.2 H50
keep ADP riding C55 1.2 H68 H67
keep hydro triang N9 2 1 C28 P1 0.87 H54
keep ADP riding N9 1.2 H54
keep geom rigid C28 O6 C24 F2 F4 F5

68 Jana2006 Cookbook, version October 2017

 keep hydro tetrahed C13 2 2 C16d N8 0.96 H36 H37
keep hydro tetrahed C15 2 2 N8 C16c 0.96 H40 H39
keep hydro tetrahed C55 2 2 C16b N7 0.96 H68 H67

OK
Open the page "Basic"
Set number of cycles to 0; OK; YES+START to save commands and start refinement
This will apply keep commands without any further changes
Start "Plot structure"
Verify the structure after application of the new keep commands
There should be only small changes
Start "Refine Commands"
Change number of refinement cycles back to 100
OK; YES+START to save commands and start refinement
The refinement converges with Rp~2.6%, R~7.8% and GOF~2.4.
Start "Plot structure"
The resulting structure has much better geometry of phenyl groups
However, the distance C13-N8 is too large

Make backup copy pfphenyl_rigid

8. Releasing the geometry of COCF3

Start "Refine commands"
In page "Various" start "Keep commands"
Delete or disable the "keep geom rigid" command; OK
Start "Distance restrains"
Define the following restrains (with default s.u.):
"Keep a value equal for all items": C24 F5; C24 F2; C24 F4

69 Jana2006 Cookbook, version October 2017

 "Fix to the specified value": C24 C28 to 1.526
"Fix to the specified value": C28 N9 to 1.35

OK; OK. YES+START to save commands and start refinement

The refinement converges with Rp~2.5%, R~7.7% and GOF~2.3.
Start "Plot structure"
The geometry is correct and (as a side effect) the distance N8-C13 is now correct
Make backup copy as phenyl_rigid_final

9. Refinement of structure and profile parameters together

Up to now the profile parameters were fixed on values obtained in Le Bail refinement
Start "Edit profile"
Activate refinement of a, b, c, beta, GW, GV, LY, HpS/L and shift
OK; YES to save changes
Start "Refine"
The refinement converges with Rp~2.5%, R~7.5% and GOF~2.3.
Start "Edit profile"
Go to page Corrections
Change number of Legendre polynomial terms to 20
OK; YES to save changes
Start "Refine"
Now the manual background is combined with the polynomial background
The problem with the background refinement, which we have seen at the beginning of this
example, cannot occur with Rietveld refinement because the intensities of the profile points
are calculated from the structure factors
The refinement converges with Rp~2.4%, R~6.2% and GOF~2.2.

10. Fixing planarity of phenyl

Start "Plot structure"
The structure is correct except that the phenyl rings are not perfectly planar
Start "Profile viewer"
The profile is correctly described
Start "Refine commands"
In page "Various" start "Keep commands"
Define "keep - geometry - plane" for C10, C11, C12, C16, C22 and C29, i.e. for the carbon
atoms of the model molecule
OK; OK; YES+START
The refinement converges with Rp~2.4%, R~6.4% and GOF~2.4
Start "Plot structure"
The phenyl rings are now perfectly planar, with reasonable distances and angles, although
only the planarity has been kept.

70 Jana2006 Cookbook, version October 2017

Plot with default radii Plot with radii scaled by isotropic ADP

71 Jana2006 Cookbook, version October 2017

72 Jana2006 Cookbook, version October 2017
Example 2.5: Twoph
Multiphase powder refinement
Revised: 5 October 2015

Room temperature (RT) phase of this sample is known from single crystal data. At cooling, a
phase transition occured which repeatedly destroyed the single crystals. Therefore, the low
temperature (LT) phase was measured at 150K as a powder. The powder sample at 150K still
contained an admixture of the room temperature phase.

Unit cell of the LT phase was determined with CRYSFIRE 2004. Its structure was solved by direct
methods implemented in the program FOX, using the RT structure as a starting template.

Input files:
twoph.xrdml with powder profile
RT.cif crystal structure of the RT phase solved and refined from single-crystal
LT.cif crystal structure of the LT phase solved from powder data by FOX. Not refined yet.

Additional information:
Powder measurement was done with Empyrean (PANalytical) using CuKα radiation, focusing
mirror and Debye-Scherrer geometry (sample in a capillary)
RT phase unit cell: 6.81 8.56 17.78 90 96.8 90
RT phase symmetry: P21/n
LT phase unit cell: 8.54 17.68 7.15 90 111.25 90
LT phase symmetry: P21/a

Chemical scheme of RT and LT:

1. Creating new job name

Start Jana2006
“File → Structure → New” opens the file manager
Left pane: locate directory with input files
Right pane: double-click twoph

2. Import procedure
[On the screen: Specify type of the file to be imported]
Select “Powder data – various CW formats”; NEXT

73 Jana2006 Cookbook, version October 2017

 [On the screen: Powder data from:]
Select "PANalytical XRDML"
Select “Debye-Scherrer method”; NEXT
[On the screen: Complete/correct experimental parameters]
For “Cell parameters” type “6.81 8.56 17.78 90 96.8 90” (cell parameters of the RT phase)
Check the radiation (should be interpreted from the xrdml file as Copper with doublet):
NEXT; FINISH; OK to close data repository; YES to accept the data set
[On the screen: question how to continue]
Select “No, I shall go by my own way”; OK

3. Powder profile indicates a multiphase

Start “Tools → Powder → Make Le Bail”
YES to start profile viewer
Press “Default” and move the slider to beginning of the powder profile
Some peaks, for example at 12.22°, do not have corresponding Bragg position. This is the
indication of a multiphase (we believe our unit cell)
Quit

4. Importing of the second phase

Start "Tools → Phases → New phase"
[On the screen: New phase]
Select "Manually"; OK
[On the screen: Define/ modify basic structural parameters for phase: Phase#2]
For “Cell parameters” type “8.54 17.68 7.15 90 111.25 90” (unit cell of the LT phase);
OK; YES to rewrite M50
In the lower left corner of the Jana2006 basic window two new thumbnails appear:
"Phase#1" and "Phase#2". They are used for switching between the two phases. Many tools
of Jana2006 work only with the active phase, for instance “Fourier”, “Dist”, “Contour”,
“Graphics viewer”. Before starting these tools, the required phase must be activated. Some
other tools like “EditM50” and “Edit profile” can switch the phase internally. "Refine" and
"Profile viewer" work independently on the active phase, and the tool "Structure solution"
places results to the first phase and deletes atoms of the other phases.

5. Set Phase's labels and symmetry

Start "EditM50"
[On the screen: Define/ modify basic structural parameters for phase: Phase#1]
Ensure that Phase#1 is selected in the bottom of the EditM50 window.
For “Phase label” type "RT"
Still in the EditM50 window, switch to the Phase#2
For “Phase label” type "LT"
Go to the page "Symmetry"
Select the phase RT and for “Space group” type “P21/n”; TAB
Select the phase LT and for “Space group” type “P21/a”; TAB
OK; YES to rewrite M50

6. Le Bail refinement
Start "Edit profile"
[On the screen: Powder options for phase: RT]

74 Jana2006 Cookbook, version October 2017

 Ensure that the RT phase is selected at the bottom of the form.
In page “Cell” activate refinement of a, b, c and beta (RT phase)
Select LT phase
In page “Cell” activate refinement keys of a, b, c and beta (LT phase)
Go to the page “Profile”
Select RT phase
Select “Pseudo-Voigt” and activate refinement keys for GW and LY
Select LT phase
Select “Pseudo-Voigt” and activate refinement keys for GW and LY
Go to the page "Corrections"
Corrections are independent on a phase selection
For “Background” select “Legendre polynomials” and for "Number of terms" type 25
OK; YES to rewrite changes

Right-Click on “Refine”
[On the screen: Refine commands]
Set "Number of cycles" to 30
OK; Yes+start
Refinement converges with GOF~3.3%, Rp~1.27%, 37 profile parameters

7. Correction for asymmetry

Start “Profile viewer” and investigate the powder profile
The peaks at low angles are slightly asymmetric, otherwise the profile fits very well
Note two rows of Bragg positions
Quit the profile viewer
Start "Edit profile"
[On the screen: Powder options for phase: RT]
Go to "Asymmetry/Diffractometer"
Select "correction by divergence"
Enable refining of HpS/L , set HmS/L to 0 and do not refine it.
Go to "Corrections"
Check "shift"
OK; YES; Start Refine
R factors will fluctuate during the refinement.
In the refinement progress window, press “Parameters” and change the Damping factor to
0.5. OK to continue refinement.
Refinement converges to final Rp~1.2%
Waiting for complete convergence is not necessary

Start "Profile viewer"

Select by mouse the range between 9°2Theta to 19°2Theta

75 Jana2006 Cookbook, version October 2017

 There are two rows of vertical bars. Clicking on some of them gives information about h k l,
2Theta position and about the corresponding phase. In this case, the top row corresponds to
the LT phase while the bottom row corresponds to the RT phase.
Quit the profile viewer

8. Backup copy
The Le Bail fitting is now complete, create a backup copy!
Start “File → Structure → Save as"
Make a backup copy "twoph_lebail"
OK; NO (Do not continue with the new structure)

9. Importing structure models for RT and LT

Note: in this example we could import structures of both phases already at the beginning
and make Le Bail fit with “Make only profile matching” refinement option. By this way we
demonstrate that structures can also be inported later on, e.g., in case when we need to test
more structure models
Structure models cannot be imported through “File → Import model from” because this
would destroy the phases.
In next steps, we rely that the backup copy created in the previous step exists!
Start “Tools → Phases → Delete Phase” and delete the phase LT
Start “File → Structure → Import model from → CIF” and import the file “RT.cif”
The program retains the original unit cell, i.e. the one refined by the previous Le Bail fitting.
This is important the unit cell in RT.cif corresponds to the room temperature measurement
while this data set has been acquired at 150K
Start “Tools → Phases → New Phase”
[On the screen: New phase”]
Select “Import from CIF format” and import the file “LT.cif”

76 Jana2006 Cookbook, version October 2017

 Start “EditM50” and rename “Phase#1” to “RT”, “Phase#2” to “LT”
Start “File → Structure → Copy In” and copy only M41 from the backup copy
"twoph_lebail"
Start refinement option, page “Basic”, and check “Make only profile matching”
Start refinement.
There should be only small changes and quick convergence to Rp~1.2%
Start “File → Structure → Save as"
Make a backup copy "twoph_rietveld_start"
OK; NO (Do not continue with the new structure)

10. Rietveld refinement of both phases with fixed parameters

In this example, we will use combination of rigid body with bond and bond-angle restraints.
The rigid body approach will be applied on the RT phase because (i) this phase was known
from single-crystal measurement and (ii) this is the minor phase. In fact, we are refining the
RT phase only to get better results for the LT phase.

Start "Edit profile"

[On the screen: Powder options for phase: RT]
Uncheck a, b, c and beta for both phases
Uncheck GW and LY for both phases
Uncheck asymmetry correction
Go to "Corrections"
uncheck shift
OK; YES

Right-Click on Refine
[On the screen: Refine commands]
Set "Number of cycles" to 100
Uncheck "Make only profile matching"
Go to "Various" and open "Fixed commands"
Fix all structure parameters (“*” for Atoms/parameters; “Add”)
OK; OK; Yes+start
Refinement converges with Rp ~ 6%

11. Keep commands for hydrogen atoms of the LT phase

[On the screen: Basic window of Jana2006]
Select LT phase
Go to "Edit Atoms"
Note: Be sure you are editing the LT phase!
Select all hydrogen atoms and delete them.
Select all carbon atoms, N2_2 and N3_2
N1_2, Cl1_2 and Cl2_2 should not be selected
Click "Action" - > "Adding of hydrogen atoms"; OK
[On the screen: Adding "hydrogen" atoms for "N2_2"]
Ensure that geometry is "Trigonal" and neighbors are (C7_2, N1_2); "Apply->Go to
ext"
[On the screen: Adding "hydrogen" atoms for "N3_2"]
Ensure that geometry is "Trigonal" and there is only one neighbor C7_2; "Apply "
77 Jana2006 Cookbook, version October 2017
 OK; YES
By this we have generated fixed commands for hydrogen atoms. Total number of atoms
should be 21

12. Rigid body for the RT phase

Right-Click on Refine
[On the screen: Refine commands]
Go to the page "Basic"
Set Damping factor to 0.1
Go to the page “Various”; start "Keep commands"
Select "Geometry" and "Rigid" and click "Select atoms" button.
[On the screen: "Select atoms to define rigid body unit"]
Select all atoms of the RT phase (atoms without “_2”); OK; Add
The new keep command has the following form: “keep geom rigid Cl1 Cl2 S1 N1 N2 N3 C2 C3
C4 C5 C6 7 C8 C9 H1c2 H1c4 H1c6 H1c9 H1n3 H1n2 H2n3”
OK;

13. Set bond and bond-angle restraints

In the next we will use pre-prepared “distfix” and “anglefix” commands based on the
geometry of the RT phase. RT phase was solved from single crystal; hence, we suppose that
its geometry is correct. The format of the commands is: key word, value of the restraint,
sigma (connected to the weight of the restraint in the refinement), list of atoms

[On the screen: Refine commands]

Go to "Distance restraints"
Click "Edit" to open the text file editor.
Copy (from distfix.txt) and Paste these lines:
distfix 1.736 0.001 C5_2 Cl1_2
distfix 1.401 0.001 C3_2 C5_2
distfix 1.376 0.001 C4_2 C5_2
distfix 1.460 0.001 C2_2 C3_2
distfix 1.400 0.001 C8_2 C3_2
distfix 1.369 0.001 C6_2 C4_2
distfix 1.265 0.001 N1_2 C2_2
distfix 1.741 0.001 Cl2_2 C8_2
distfix 1.375 0.001 C9_2 C8_2
distfix 1.371 0.001 C9_2 C6_2
distfix 1.371 0.001 N2_2 N1_2
distfix 1.345 0.001 C7_2 N2_2
distfix 1.685 0.001 S1_2 C7_2
distfix 1.313 0.001 N3_2 C7_2
Close text file editor and save changes
OK
[On the screen: Refine commands]
Go to "Angles restraints"
Click "Edit" to open the text file editor.
Copy (from anglefix.txt) Paste these lines:
anglefix 121.977 0.01 C3_2 C5_2 Cl1_2
78 Jana2006 Cookbook, version October 2017
 anglefix 115.676 0.01 C4_2 C5_2 Cl1_2
anglefix 122.347 0.01 C4_2 C5_2 C3_2
anglefix 126.236 0.01 C2_2 C3_2 C5_2
anglefix 114.761 0.01 C8_2 C3_2 C5_2
anglefix 120.262 0.01 C6_2 C4_2 C5_2
anglefix 118.989 0.01 C8_2 C3_2 C2_2
anglefix 124.469 0.01 N1_2 C2_2 C3_2
anglefix 120.003 0.01 Cl2_2 C8_2 C3_2
anglefix 123.439 0.01 C9_2 C8_2 C3_2
anglefix 119.969 0.01 C9_2 C6_2 C4_2
anglefix 116.048 0.01 N2_2 N1_2 C2_2
anglefix 116.553 0.01 C9_2 C8_2 Cl2_2
anglefix 119.211 0.01 C6_2 C9_2 C8_2
anglefix 119.412 0.01 C7_2 N2_2 N1_2
anglefix 119.203 0.01 S1_2 C7_2 N2_2
anglefix 116.961 0.01 N3_2 C7_2 N2_2
anglefix 123.835 0.01 N3_2 C7_2 S1_2
OK

14. ADP for the RT and LT phase

[On the screen: Refine commands]
Start "Fixed commands"
Delete the "fixed all *" command created previously
Select “ADP harmonic parameters”
For “Atoms/Parameters” type “Cl1 Cl2 S1 N1 N2 N3 C2 C3 C4 C5 C6 C7 C8 C9 H1c2 H1c4
H1c6 H1c9 H1n3 H1n2 H2n3”
Add; OK
This fixes ADP of all atoms of the RT phase to the values refined from the single crystal data

[On the screen: Refine commands]

Go to "Restrictions"
Set one ADP parameter identical for all non-hydrogen atoms of LT phase
Uncheck "Make identical coordinates"
Check "Make identical ADP parameters"
For occupancies use “not restricted”
For “Atoms/Parameters” type “C5_2 C4_2 C6_2 C9_2 C8_2 C3_2 C2_2 N1_2 N2_2 C7_2
S1_2 N3_2 Cl1_2 Cl2_2”
Add; OK
Using this restriction, the program will refine one ADP common for all atoms of the LT phase

OK; Yes;
Make a backup copy (File->Structure->Save as)
Start refinement
Press "Use all" as a reply to the message about significant dicsrepancy
Refinement converges with R~4.7% (RT), 4.3% (LT), Rp ~ 3%

Plot both structures by "Plot structure" and check them.

79 Jana2006 Cookbook, version October 2017

 Check also the profile fit in "Profile viewer"

15. Fine tuning the profile parameters

Start "Edit Profile"
[On the screen: Powder options for phase: RT]
Select the LT phase
Activate refinement of a, b, c, and beta
Go to the page "Profile"
Activate refinement of GW and LY of LT phase
Go to the page "Corrections"
Activate refinement of shift
Go to the page "Sample"
For Preferred orientation select “March-Dollase”
Set “pref1” to 1, activate its refinement and for “Direction” type “0 1 0”
This corresponds to the longest unit cell parameter of LT
Still in the Sample page, switch to the RT phase
For Preferred orientation select “March-Dollase”
Set “pref1” to 1, activate its refinement and for “Direction” type “0 0 1”
This corresponds to the longest unit cell parameter of RT
OK; YES (to save profile parameters)
Start refinement
Refinement converges with R~4.0% (RT), 3.6% (LT), Rp ~ 2.6%

16. Final model

Finally, we will remove ADP restrictions for heavy atoms of the LT phase
Right-click on Refine
[On the screen: Refine commands]
Go to the page "Various" and start "Restrictions"
Load the restriction command, delete Cl1_2, Cl2_2 and S1_2, rewrite the command
OK; OK; Yes+start
Refinement converges with R~3.7% (RT), 3.3% (LT), Rp ~ 2.5%
In case the convergence is too slow, Marquard rerefinement can be activated by pressing
“Parameters” during the refinement. In such case damping factor can be increased to 0.5.

80 Jana2006 Cookbook, version October 2017

Note about cell parameters of the RT phase

The CIF for the RT contains unit cell determined at room temperatire. However, our powder
pattern was measured at 150 K. The unit cell parameters are changing with temperature and it
may happen that the change it so large that refinement from the room temperature unit cell
does not converge.

We can estimate the deviation of the RT unit cell from the powder profile before its refinement:

Right-click on Refine
[On the screen: Refine commands]
Set Number of cycles to 0;
Check "Make only profile matching" to perform only Le Bail refinement.
OK; Yes+start; (This calculate theoretical pattern)
Go to the "Profile viewer"
Display the range from 9° to 19° 2Θ.

We can see that while for LT Bragg positions fit well, RT Bragg positions (second line in the plot)
are deviated, e.g., 0 1 1, -1 0 1, 0 1 3.

For rough estimation of the proper starting point for refinement of the RT unit cell, we can
assume that beta does not change with temperature and we can calculate a, b, c from
observed positions of d011= 7.7 Å, d013 = 4.85 Å and d-101 = 6.58 Å: a = 6.84 Å, b = 8.55 Å, c =
17.77 Å.

For obtaining d, use “Draw in d’s”, find reflection hkl by “Go to” button and find its real position
by “Info” line (it must stay a while at the peak position).

81 Jana2006 Cookbook, version October 2017

82 Jana2006 Cookbook, version October 2017
Example 3.1: AD3
Simple structure with pseudo-merohedric twinning and unequal twin volume
fractions. Finding twinning matrix by Symmetry wizard. Creating publication CIF.
Revised: 18 April 2017

Bis[N-(2-benzylidenepropylidene)phenyl]ether

Single crystal data measured with Oxford Diffraction four-circle diffractometer

Input files: AD3.hkl, AD3_red.sum
Frame scaling, absorption correction: done with software of diffractometer

1. Creating new jobname

Start Jana2006
“File → Structure → New” opens a file manager
Left pane: locate directory with input files
Right pane: double-click AD3

2. Import Wizard
Select “Known diffractometer formats”; NEXT
Select “Oxford Diffraction → CCD”; NEXT
Select “Input from “sum” file”; OK
Leave all settings unchanged; NEXT
Leave all settings unchanged; NEXT
For absorption correction select “None or done before importing”; NEXT
FINISH

3. Data Repository
The line in data repository window corresponds to one data set.
OK; YES to accept the data set

4. Symmetry Wizard, twinning

[On the screen: “Tolerances for crystal system recognition”.]
Select "Introduce twin laws in case of subgroups"
Leave all other settings default; NEXT
[On the screen: Select Laue symmetry]
Select Monoclinic – setting “c”
The highest Laue symmetry consistent with cell parameters is orthorhombic. The
83 Jana2006 Cookbook, version October 2017
 corresponding Rint is rather high (~17%), and this indicates violation of the orthorhombic
symmetry. Typical reason for such violation is the presence of twinning: structure of the
twins has a lower symmetry than is the highest Laue symmetry, and the twin domains are
related by a symmetry operation from the highest Laue symmetry, which is not present in
the structure of twins. With equal domain fractions, such twin could simulate an
orthorhombic diffraction pattern with very good Rint. Here the Rint is high, and this
indicates that twin domain fractions are not equal.
The selected Laue symmetry is monoclinic. In the previous step, we have activated an
option "Introduce twin laws in case of subgroups". This means that the symmetry
operation generating the orthorhombic symmetry from the selected monoclinic
symmetry will be used as the twinning operation.
NEXT
Select primitive unit cell; NEXT
[On the screen: Select space group]
Using the criteria described in the first example select P21/n, press “Details” to verify it is
really acceptable
The 46 reflections violating the extinction rules of P21/n have mostly low I/ρ. Some
additional spots in the diffraction pattern, e.g. from a small crystal attached at the surface
of the measured sample might cause the violations. Without seeing the CCD frames we
cannot be sure that the 46 reflections are really spurious, but P21/n is the most probable
symmetry.
NEXT
Select "Accept twinning matrices induced by the space group test"
Select "Accept the space group in the standard setting"; FINISH

5. Creating refinement reflection file

NEXT to confirm threshold 3sigma
[On the screen: settings for merging of reflections]
Leave all settings default; NEXT
Accept the new DatBlock and calculate coverage; FINISH

6. Structure Solution Wizard

[On the screen: window of Structure solution]
Solution of twins does not work so well like for non-twinned crystals. Therefore, we will
make several attempts and let Superflip to select the best result.
In “Formula” textbox type list of chemical elements for AD3: C N O H
Select “Superflip”, “Peaks from Jana2006”
Check "Repeat Superflip: Number of runs" and 15 for number of runs; limit Maxcycles to
1000;
Check “Use local normalization”
leave other settings default

84 Jana2006 Cookbook, version October 2017

 Press “Run Solution”
[On the screen: listing of Superflip]
With this setting, Superflip repeats calculation 15 times (from different randomly
generated set of phases) and returns the result with the best fit between the electron
density map and the user-defined symmetry. Because of twinning the fit of symmetry
generators is worse and it may happen that Superflip suggests P21 instead of P21/n -
please ignore this fact, "n" is confirmed by systematic extictions. With this structure, the
results of Superflip are significantly better when "Biso" is set to 3 in the charge flipping
options; however, this would be an empirical step useless for other structures.
Close listing of Superflip
Press “Draw structure” to see result of charge flipping
It may happen that Diamond shows two molecules instead of one. This is because
twinning also biases the bond lengths, some distances may be too long and some atoms
improperly assigned
Quit Diamond without saving changes
In case of doubts repeat structure solution and plotting
Finally press “Acccept last solution” to close the structure solution wizard

7. Verification of the structure solution

Start “Edit atoms”
[On the screen: list of atoms]
Press “Select all”
Start “Action → Edit/define”
Double-click “Type” textbox
Double-click unlocks the “Type” textbox. This is locked for security reasons because the
selected atoms have different chemical types.
Select Carbon, OK
[On the screen: list of atoms. The labels are still the original ones but the chemical type of
all atoms is carbon]
Start “Action → Rename selected atoms to atom type + number”

85 Jana2006 Cookbook, version October 2017

 Press “Select all”
Start “Action → Make symmetrically contiguous motifs”
OK; YES to save changes
These steps help to get a reasonable plot in Diamond (next step) even with biased
distances and improper chemical types
Start “Plot structure”
Press “Draw+continue”
In this mode Jana will launch Diamond but it will not wait for its end
[On the screen: window of Diamond]
Start “Build → Get molecules” or use on the bottom toolbar
Rotate the molecule ( and on the bottom toolbar) to get an optimal view
Draw rectangle to select all atoms of the molecule or press Ctrl-A
Right-click on one of the selected atoms, choose “Add → Atom labels”
[On the screen: dialogue of Diamond for atom labeling]
For “Content” select: “Atom symbol”
Define relative position of labels 0.3,0.3,10
Define text size ~0.8Å; OK
Rotate the molecule to similar orientation like has the chemical scheme
[On the screen: plot of the molecule with atom labels]

The results may be slightly different for each run of Superflip. This is because Superflip
starts from random phases and the resulting electron density map differs for each run.

Compare the plot with the chemical scheme and note labels of atoms, which must be
oxygen or nitrogen. In our example, C2 and C3 should be nitrogen and C1 should be
oxygen.
Quit Diamond

8. Assignment of correct chemical types

[On the screen: basic Jana window]
Start “Edit atoms”
[On the screen: list of atoms]

86 Jana2006 Cookbook, version October 2017

 Select the atom which should be oxygen (C1 in our example)
Editing of one atom can be also started by right mouse button or by double-click
Start “Action → Edit/define” and choose chemical type “O”; OK
Press “Refresh”
Hold Ctrl and select atoms which should be nitrogen (C2,C3 in our example)
Start “Action → Edit/define” and choose chemical type “N”; OK
[On the screen: list of atoms, labels are still the original ones but chemical types are
correct]
Press “Select All”
Press “Action” or right-click on one of the selected atoms
Choose “Rename selected atoms to atom_type + number”
[On the screen: list of atoms with labels corresponding to chemical types]
You should see one oxygen, two nitrogens and carbons. If not press ESC and start again 

OK; YES to save the changes

Check chemical types in the structure plot with Diamond

9. Refinement
[On the screen: basic window of Jana]
Right-click the icon of Refine.
[On the screen: refinement options]
Define 100 of cycles, Instability factor 0.02; OK
Choose “YES+START”
Refinement converges with R value about 16%, GOF about 8, 141 refined parameters

10. Twin volume fraction

Start “parameters → Twin fractions”
The volume fraction of the twin domain introduced by the Symmetry wizard is 0.5 and
twinning matrix relating the monoclinic and orthorhombic symmetry is (1 0 0 | 0 1 0 | 0 0 -
1).
Activate refinement of the twin fraction
Repeat refinement
Refinement converges with R value about 11%, GOF about 7
Twin volume fraction drops to 0.29.
Start “Edit atoms”
[On the screen: list of atoms]
Press “Select all” and “Action → Edit/define”
Choose “harmonic” for ADP parameters. This defines anisotropic temperature parameters
for all atoms. OK, OK, YES to save changes
Double-click the icon Refine.
Refinement converges with R value about 9%, GOF about 6, 317 refined parameters.

11. Adding of hydrogen atoms

[On the screen: basic window of Jana]
Start “Edit atoms”
In the textbox type “C*”; TAB.
This will select atoms starting with C, i.e. carbons
Start “Action → Adding of hydrogen atoms”

87 Jana2006 Cookbook, version October 2017

 [On the screen: Adding of hydrogen atoms dialogue]
Leave all settings default; OK
[On the screen: list of atoms with new hydrogens]
Total number of atoms should be 63 (can be verified using “Select All”)
The hydrogens are kept in ideal geometry by commands which can be viewed or edited in
“Refinement commands → Various → Keep commands”.
OK to close “Edit atoms”; Yes to save changes
“Plot structure” to check results of automatic hydrogen assignment
Make refinement of the final structure
Oscillation may occur during the refinement. In such case, do not break refinement, press
Parameters, change Damping factor to 0.5 and close the dialogue. Refinement will
continue with damping.
Refinement converges with R value 3.7%, GOF 2.7, 317 refined parameters.
12. Creating CIF file
The information for the CIF file is cumulated in the file M70 during the data processing as
well as solution and refinement steps. For creating a publication CIF, repetition of the basic
steps is recommended in order to refresh the information in M70. In our case, we have
started from the *.sum file so M70 does not contain any information about the
experiment.
Start “File → Reflection file → Create refinement reflection file”
In our case this step is not necessary because it has been already done

Right-click the icon of Refine; check “Refinements on F(obs)**2”.

Refinement on F2 is standard and requested by many editors
Repeat refinement

Right-click Fourier; select “F(obs)-F(calc) - difference Fourier”; make sure that “Weighting
of reflections” is checked and “Apply sin(th)/lambda limits” unchecked; in the page
“Scope” press “Reset to default”; in the page “Peaks” select “in a fixed sphere of radius
0.8”
OK; YES+START; NO to include new atoms; NO to see the listing

Right-click Dist icon

Check “d(max) derived from atomic radii and typical distances” and adjust “expanded by”
value” (depends how many bonds should be calculated); check “Calculate angles”; if
needed, adjust limits for H-bonds;
In the page “Select atoms” select central atoms and neighbor atoms – usually we want to
have all atoms in both roles;

88 Jana2006 Cookbook, version October 2017

OK; YES+START

Start “File → CIF utilities → Make CIF file”

Note about goodness of fit

The instability factor calculated from the reflection statistics is about 0.001 (see Edit/View
→ view of Reflection report, search for the string “instability”). Application of this factor in
refinement (Refine options → page Basic → “Instability factor from reflection statistics”)
leads to GOF about 6. This indicates there is some systematic error in data. For accurate
structure study, we would need to re-measure or re-process data and obtain a better data
set.

89 Jana2006 Cookbook, version October 2017

90 Jana2006 Cookbook, version October 2017
Example 3.2: PyNinit
Simple structure with non-merohedric twinning. Handling twin overlaps in
Jana2006.
Revised: 18 April 2017

Acta Cryst. (2009). C65, m260-m262

Single crystal data measured with Oxford Diffraction four-cycle diffractometer

Input files:
1st twin domain: pyNinit_twin1.hkl, pyNinit_twin1_red.sum
2nd twin domain: pyNinit_twin2.hkl, pyNinit_twin2_red.sum
hklf5: pyNinit_twin1_hklf5.hkl
Twinning matrix: (-1 0 -0.733| 0 -1 0| 0 0 1)
Frame scaling, absorption correction: done with software of diffractometer

This figure shows diffraction pattern of pyNinit – peak positions projected along b* with unit
cells of both twins related by the above-mentioned twinning matrix
91 Jana2006 Cookbook, version October 2017
1. Creating new jobname
Start Jana2006
“File → Structure → New” opens a file manager
Left pane: locate directory with input files
Right pane: double-click pyNinit_twin1

2. Import Wizard (for the 1st twin domain)

Select “Known diffractometer formats”; NEXT
Select “Oxford Diffraction-CCD”; NEXT
Select “Input from “sum” file”; OK
Leave all settings unchanged; NEXT
Leave all settings unchanged (leave Twinning checkbox clear); NEXT
Information: All 34630 input reflections were properly handled; OK
For absorption correction select “None or done before importing”; NEXT
FINISH

3. Data Repository
The line in data repository window corresponds to one data set.
OK; YES to accept the data set

4. Symmetry Wizard
[On the screen: an information that symmetry wizard follows]
NEXT
[On the screen: “Tolerances for crystal system recognition”.]
Leave all settings default; NEXT
[On the screen: Select supercell]
Select continue with the basic cell; NEXT
[On the screen: Select Laue symmetry]
Select Monoclicnic; NEXT
[On the screen: Select cell centering]
Select primitive unit cell; NEXT
[On the screen: Select space group]
Using the criteria described in the first example Select P21/c; NEXT
[On the screen: Final step of the space group test]
Accept the space group transformed into the original cell; FINISH
Follow the wizard for creation of the refinement reflection file

5. Structure solution
[On the screen: window of Structure solution]
In “Formula” textbox check list of chemical elements for PyNinit: Ni C N O H
Select “Superflip”, “Peaks from Jana2006”; leave other settings default; “Run Solution”
Superflip converges (after noise suppression) with R value ≈ 23%.
Fit of symmetry elements is very good for c, slightly worse for 21
Press CLOSE to leave the listing
Press “Draw structure” to see the result
Press “Accept last solution” to close the Structure solution wizard

6. Completing the structure solution

92 Jana2006 Cookbook, version October 2017
 In Refinement options set 100 refinement cycles, damping factor 0.5
Run Refine
Refinement converges with R value about 19%.
Serious warnings appear in the listing.
Use “Edit atoms” and “Select rejected” to delete atoms which may be automatically
rejected during the refinement because of too large Uiso
(The button is disabled when there are no such atoms)
Using “Plot structure” identify the incorrectly assigned chemical types by comparison of the
plot with the chemical scheme
Using “Edit atoms” correct the chemical types, delete wrong atoms, “rename atoms
according to chemical type” and “create symmetrically contiguous motifs”
Repeat refinement
Run difference Fourier and complete the possibly missing atoms
Repeat refinement
Refinement converges with R value about 17%.
Using “Edit atoms” set harmonic ADP for all atoms
Refinement converges with R value about 17%.
Serious warnings appear in the listing duse to non-positive definite ADP of some atoms
Using “Edit atoms” add hydrogens to carbon and nitrogen atoms
For nitrogen atoms skip cases when Ni is listed as a neighbor (this is the nitrogen
coordinated to Ni) and skip also cases when nitrogen has three oxygen atoms as neighbors
(this is NO3 group).
Check the hydrogen atoms by plotting. The geometry is slightly distorted due to unresolved
twinning which may confuse the automatic procedure for adding hydrogens.
Finally, the structure model should have 53 atoms
Repeat refinement
Refinement converges with R value around 16.6% and GOF 9, 316 parameters
Several non-positive definite ADPs are indicated in the List of serious warnings in the
refinement listing

7. Twinning
In “EditM50” check “Twinning” and define twinning matrix (-1 0 -0.733| 0 -1 0| 0 0 1)
The twinning matrix is known from the diffractometer software and represents a rotation by
180º around c. The program offers re-creation of the refinement reflection file because
twinning may change rules for merging of symmetrically equivalent reflections. In this case,
it is not necessary.
In “Parameters → Twin fractions” enable refinement of “twvol2”
Repeat refinement
Refinement converges with R value around 13% and GOF 8 for 317 parameters
Open listing of Refine
ADP tensors are still not positive definite for some atoms.

This twin has a diffraction pattern where some reflections are partially overlapped. In
refinement options there are two limits available for treating partial overlaps. The
reflections with difference in theta smaller than “Maximal angular difference for twin
overlap” are treated as fully overlapped; the reflections with theta difference larger than
“Minimal angular difference for full separation” are treated as fully separated; the
reflections between the limits are discarded. The limits should be adjusted carefully not to
93 Jana2006 Cookbook, version October 2017
 discard too many reflections.

In Refinement options change values of “Maximal angular difference for twin overlap” (L1 in
the table) and “Minimal angular difference for full separation” (L2 in the table) and watch
their impact on R factors, number of excluded reflections and ADP tensors. Find an optimal
combination.

(This table is calculated for refinement on F and instability factor 0.01)

L1 L2 Robs GOF discarded
0 0 15.3 8.2 0
0.1 0 12.8 7.4 0
0.2 0 7.4 3.6 0
0.3 0 7.1 3.5 0
0.4 0 9.8 5.5 0
0.35 0 9.6 5.4 0
0.31 0 9.3 5.3 0
0.3 0.4 6.9 3.6 778
0.3 0.5 7.4 3.8 1264
0.3 0.6 7.9 4.1 1916
0.2 0.3 6.2 3.0 415
0.2 0.4 5.7 3.0 1193
0.15 0.3 5.3 2.4 709
0.1 0.3 4.8 1.9 1120
0.1 0.4 3.9 1.4 1898

With maximal angular difference for twin overlap 0.1 and minimal angular difference for full
separation 0.4, refinement converges with R value 3.9% and GOF 1.4, all ADP tensors are
positive definite, the fraction of the second twin domain refines as 0.468(2) (see
“Parameters → Twin fractions)”.
Using “Structure → Save as”, create a backup copy, for instance pyNinit_1domain

8. Import data of the second domain

Usually (and also in this case) the first twin domain + twinning matrix is sufficient for the
refinement. If data of the second domain are good, we can add it as another data set to get
better refinement of the twin fraction.
“File → Reflection file → Import/modify the reflection file”; Import new
Known diffractometer formats; Oxford Diffraction-CCD
Browse for PyNinit_twin2.hkl
Select “Input from “sum” file”; OK ; OK
Leave all settings unchanged; NEXT
[On the screen: Relationship to the reference cell/split by twinning]
Here we have to define that reflections of the second domain were indexed according to
orientation matrix of the second domain.
Data related to domain #2; NEXT; Accept cell parameters
[On the screen: Information about difference in cell parameters]
Select “Continue with the old reference cell parameters”

94 Jana2006 Cookbook, version October 2017

 YES to suppress possibly duplicate reflections
Data processing of each twin was done on the same experimental frames. Therefore,
overlapped reflections are processed twice with the same result. Merging of duplicitely
processed reflections would lead to very low sigma and unrealistic high weight in the
refinement.
For absorption correction select “None or done before importing”; NEXT
FINISH
[On the screen: Data repository]
Each line in data repository window corresponds to one data set.

OK; YES to accept the data set;

In this step the program offers sorting of the data sets to data blocks. Each data block has
an independent scale factor. In our case, both twin domains come from the same
measurement and they should be used as one data block.
Set Data collection block #1 for both datasets
Follow the wizard for creation of the refinement reflection file; select “on the same scale”
for both data sets
Refine
Refinement converges with R 4.1%, GOF 1.36, 3800 excluded reflections, fraction of the
second domain is 0.4641(6) (the number with s.u. is printed in the refinement listing).
Create a backup copy, e.g., pyNinit_2domains

9. Are both domains really on the same scale?

It may be unclear whether both data sets are on the same scale. Different scales would bias
refinement of the volume fraction.
“File → Reflection file → Import/modify the reflection file”
In Data repository highlight pyNinit_twin2.hkl and use “Reimport”
Follow the import wizard as previously, but do NOT suppress possibly duplicate reflections
Duplicate reflections will be used to find the common scale
Follow the steps the wizard for creation of the refinement reflection file; select “on
different scales” for both data sets

95 Jana2006 Cookbook, version October 2017

 In the next step, Jana will determine the unique scale based on the duplicate reflections

The scale is calculated from 4334 pairs of equivent reflections, which should have the same
intensity in both data sets after the scaling.
Refine
Refinement converges with R about 4.1%, GOF 1.39, 3800 excluded reflections, fraction of
the second twin domain is 0.4716(6). However, in this refinement we have unrealistic weight
of the duplicate reflections because they were entered two times with the same sigma.

10. Use of HKLF5

The diffractometer software should better detect overlaps of diffraction spots than the
simple criterion in Jana2006. HKLF5 file produced by diffractometer software contains flags
saying which reflections are overlapped. Jana2006 can work with HKLF5 if twinning is
activated (twinning matrix can be arbitrary).
“File → Reflection file → Import/modify the reflection file”; Delete both datasets;
Import new
[On the screen: Specify type of the file to be imported]
Choose “reflection file corrected for LP and absorption”, NEXT
Select “SHELX hklf5”
Browse for pyNinit_twin_hklf5.hkl; NEXT
Leave all settings unchanged; NEXT
Leave all settings unchanged; do not select “HKLF5 file by Jana tools”;

NEXT;
OK to confirm information about 68907 imported reflections
FINISH
OK; YES to close and save Data repository
Follow the wizard for averaging of reflections and creation of the refinement reflection file
Jana2006 merges only fully separated reflections from both domains
Refine

96 Jana2006 Cookbook, version October 2017

Refinement converges with R 5.4%, GOF 1.6, fraction of the second domain is 0.4728(5). The
R value is slightly larger than previously but we are using all reflections. Results based on
this refinement should be considered the most reliable.

The following table summarizes refinement of the twin volume fraction

twvol2
First twin domain, L1 = 0.1, L2 = 0.4 0.468(2)
Both domains, L1 = 0.1, L2 = 0.4 0.4644(6)
Both domains, L1 = 0.1, L2 = 0.4, scaling 0.4716(6)
HKLF5 0.4728(5)

97 Jana2006 Cookbook, version October 2017

98 Jana2006 Cookbook, version October 2017
Example 3.3: CsLiSO4
Simple structure with pseudo-merohedric 3-fold twinning. Finding twinning
matrix from group → subgroup transformation. Transformation to four times
larger reciprocal cell.
Revised: 23 October 2017
CsLiSO4
Single crystal data measured with Oxford Diffraction four-cycle diffractometer
Input files: CsLiSO4.hkl, CsLiSO4_red.sum
Frame scaling, absorption correction: done with software of diffractometer

1. Creating new jobname

Start Jana2006
“File → Structure → New” opens a file manager
Left pane: locate directory with input files
Right pane: double-click CsLiSO4

2. Import Wizard
Select “Known diffractometer formats”; NEXT
Select “Oxford Diffraction-CCD”; NEXT
Select “Input from “sum” file”; OK
Leave all settings unchanged; NEXT
Leave all settings unchanged (leave Twinning checkbox clear); NEXT
For absorption correction select “None or done before importing”; NEXT
FINISH

3. Data Repository
The line in data repository window corresponds to one data set.
OK; YES to accept the data set

4. Symmetry Wizard
[On the screen: an information that the next tool is the symmetry wizard]
NEXT
[On the screen: “Tolerances for crystal system recognition”.]
Leave all settings default; NEXT
[On the screen: Select Laue symmetry]

99 Jana2006 Cookbook, version October 2017

We can see that hexagonal symmetry is slightly violated, because corresponding Rint is about
14%. On the other hand, an orthorhombic symmetry with setting “-a-b,a+b” has Rint
comparable with the triclinic symmetry, given much higher redundancy of equivalent
reflections.

For exactly hexagonal cell parameters, merohedric twinning is highly probable. Another
indication of twinning is the fact that the three equivalent orthorhombic settings have quite
different Rint. It might be caused by merging reflections belonging to different unequally
populated twin domains.

In the next step, we will repeat test of Laue symmetry but this time considering twinning.
That means for any subgroup of 6/mmm the program will generate corresponding pseudo-
merohedric twinning matrices and it will use only symmetry elements which are present in a
twinned structure for general case of not equally populated twin domains.

Press BACK
Check „Introduce twin laws in case of subgroups“

NEXT

100 Jana2006 Cookbook, version October 2017

 [On the screen: Select Laue symmetry]

This time, the three orthorhombic cases gives the same internal R factor.
Select the highlighted orthorhombic Laue symmetry; NEXT
The three orthorhombic cases are related the three twin domains. The selected one gave us
in the previous test the lowest Rint. This means that this domain is more populated than the
others.
Select C centering; NEXT
Hexagonal unit cell transformed to an orthorhombic one must be C centered
[On the screen: Select space group]
Select Ccce space group; NEXT
There are several possibilities for selecting space group and at this stage it is unclear which
is the correct one.
Accept the space group in the standard setting; FINISH
Follow wizard for creation of the refinement reflection file
QUIT the wizard for structure solution without solving the structure

5. Space group test results

[On the screen: basic window of Jana]
Check new cell parameters and symmetry with EditM50
Cell parameters 10.8919 18.8653 8.8048 90 90 90
Space group Ccce
ESC
[On the screen: basic window of Jana]
Start “Parameters → Twin fractions”

101 Jana2006 Cookbook, version October 2017

 The initial volume fractions are equal but we already know they will be different; otherwise
hexagonal Laue symmetry would be present with good Rint.
Focus the textbox “twvol2” and press “Show twinning matrix”
0.5 -0.5 0 / 1.5 0.5 0 / 0 0 1
Focus the textbox “twvol3” and press “Show twinning matrix”
-0.5 -0.5 0 / 1.5 -0.5 0 / 0 0 1
These twinning matrices are related to the current orthorhombic unit cell
Return to the basic window

Start “Edit/view → Editing of M90 file”

eformat91 0
datblock Block1
nref 2021 obslim 3 sctom90 1 norefitems 1 diffscales 1 dataave 1 scalelim
10 hklf5 0
indslowest 3 indfastest 1 addcentrsymm 0 sigimethod 1 multave 1
flimprint 5 flimcull -1 avesigwt 0 unstab 0.022791
radtype 1 dattype 1
lambda 0.71073 lpfactor 1 monangle 6.0821 perfmono 0.5
datcolltemp 293
end
Data Block1
2 0 0 166.8 3.4 1 0 1 0.0000
4 0 0 115151.2 357.3 1 0 1 0.0000
6 0 0 13.8 9.2 1 0 1 0.0000
8 0 0 56835.4 139.6 1 0 1 0.0000
10 0 0 -8.5 16.5 1 0 1 0.0000

The numbers indicated in bold are twin flags. All flags except several reflections at the end
of the file have the twin flags equal to “1” because most reflections can be indexed in the
first domain. The reflections at the end of the file with the twin flag “2” and “3” are the
ones, which cannot be transformed to the first domain because after such transformation
they would break extinction rules for the selected space group.

6. The attempt to solve the twinned orthorhombic structure

[On the screen: basic window of Jana]
“Run → Structure solution”
[On the screen: window of Structure solution]
In “Formula” textbox type list of chemical elements for CsLiSO4: Cs Li S O
Select “Superflip”, “Peaks from Jana2006”; leave other settings default; “Run Solution”
Repeat it several times. The selected space group has never been confirmed by Superflip, it
mostly suggests Cmcm.
Press Change the space group; for Space group type Cmcm; OK, YES to rewrite M50 file
Follow steps of the wizard for creation of the refinement reflection file until the program
returns to the Structure solution wizard
[On the screen: window of Structure solution]
Press “Run solution”
Repeat it several times until Superflip confirms Cmcm. In some runs, Superflip might suggest
different symmetry, which is indication of some structure solution troubles.
Press “Draw structure”; Select “Draw+Continue”;
[On the screen: window of Diamond]

102 Jana2006 Cookbook, version October 2017

 Start “Build → Get molecules” or use on the bottom toolbar
The solution looks unreasonably and other attempts to solve the structure (direct methods,
shelxt, use of different symmetries etc.) would fail as well.
Quit Diamond
Quit the Structure solution wizard (YES to discard the last solution)

7. What is wrong?
“File → Transformation → Cell transformation”
Press “to those from data collection”; OK
Select “rewrite the old structure”; OK
Follow steps of the wizard for creation of the refinement reflection file
We are now back at the original hexagonal unit cell. In EditM50, you can see there is no
symbol for the transformed symmetry operation. However, we will check for the symmetry
again.
“File → Reflection file → Make space group test”
This manually starts the Symmetry wizard.
[On the screen: Tolerances for crystal system recognition]
Check „Introduce twin laws in case of subgroups“
Select “look for centering vectors composed from 0 and ½”
This last setting is different compared with the previous symmetry check. Normally, one
looks for vectors 0, 1/3, 2/3 when the hexagonal unit cell is present, but we already know
that we will use an orthorhombic symmetry where 0 and ½ centering vectors are
reasonable.
NEXT
Select the highlighted orthorhombic Lause symmetry (like before); NEXT
[On the screen: Select cell centering]

This time, X centering is suggested

Press “Show/modify X centering” to see the individual centering vectors

103 Jana2006 Cookbook, version October 2017

We can see the centering vector (½ ½ 0) corresponding to the C centering but also additional
vectors (½ 0 0) and (0 ½ 0) suggesting that a and b unit cell parameters could be halved.
OK;
Select X centering; NEXT
Select space group Pmcn; NEXT
Accept space group in the standard setting; FINISH

YES to create refinement reflection file

Follow steps of the wizard for creation of the refinement reflection file
83 observed reflection will be discarded due to the selected space group. This is probably
caused by spurious peaks, which are present in the diffraction pattern

The transformation suggested at the final step of the space group test produces an
orthorhombic unit cell with four time smaller volume compared with the previously tested
orthorhombic unit cell. Such unit cell alone could not index all diffraction spots, but
combined with a three-fold twinning it can. The indexing of the hexagonal diffraction
pattern with a three-fold twin of this small orthorhombic unit cell is indicated in the next
figure.

104 Jana2006 Cookbook, version October 2017

8. Structure solution
“Run → Structure solution”
[On the screen: window of Structure solution]
In “Formula” textbox check list of chemical elements for CsLiSO4: Cs Li S O
Select “Superflip”, “Peaks from Jana2006”; leave other settings default; “Run solution”
Superflip converges (after noise suppression) with R value 22%. The suggested space group
by Superflip is Imma but for twins such prediction is unreliable. On the other hand, Superflip
confirms the symmetry elements of the current space group symmetry Pnma.
Press CLOSE to leave the listing
Press “Draw structure”
Press “Accept last solution”
Start EditM50; go to page Composition; change Formula units to 4; press Formula from M40
In most cases, the formula should be Cs Lin S O4, where “n” can be zero, 1 or more. This
means all heavy atoms except Li have been properly interpreted. If not, repeat the structure
solution.
Start “Edit atoms”; delete all Li atoms
“Plot structure”
Diamond should show SO4 tetrahedron and Cs. For this plot, all bonds to Cs will need to be
disabled through “Build – Connectivity”

105 Jana2006 Cookbook, version October 2017

Run refinement
“Parameters → Twin fractions”
Activate refinement of the twin fractions

Run refinement
Start “Edit atoms”; change ADP of all atoms to harmonic
Run refinement
Refinement converges with GOF ~ 4.4 and R(obs) ~ 4.4% for 36 parameters
Right-click Fourier
Select “Map type” = “F(obs)-F(calc) – difference Fourier”
OK; YES+START; YES to include new atoms
[On the screen: Inserting/replacing of atoms]
OK
[On the screen: List of peaks]

The strongest difference peak (Max1) has reasonable Li-O distances to oxygen atoms. We
can add it as Lithium in the first highlighted symmetry position.
Press “Include selected peak”
For “Name of atom” type Li1; TAB; OK; FINISH; YES to include the new atom.
Run refinement
Refinement converges with GOF ~ 3 and R(obs) ~ 2.7% for 39 parameters
106 Jana2006 Cookbook, version October 2017
YES (or Edit/View → View of Refine) to open the listing
Go to → Statistics Fo, sin(th)/lambda after refinement
The Statistics show that the strongest low-angle reflections are systematically weaker. This
is an indication that extinction correction is needed.
Close the refinement listing
Using “Parameters → Extinction” introduce isotropic Type 1 Gaussian or Lorentzian
extinction correction
Run Refine
Refinement converges with GOF ~ 2.7 and R(obs) ~ 2.2% for 40 parameters
Start “Edit atoms”; change ADP of Lithium to harmonic
Refinement converges with GOF ~ 2.6 and R(obs) ~ 2.2% for 43 parameters
Right-click Refine; go to the page Basic
Check “Instability factor from reflection statistics”
OK; YES+START
Refinement converges with GOF ~ 1.4 and R(obs) ~ 2.3% for 43 parameters
This last step has changed the weighting scheme – see Example 1.1 for details

The final structure

Right-click the icon of Refine

In the page Basic, check “Instability factor from reflection statistics”
Repeat refinement
Refinement converges with R value around 2.3% and GOF 1.4, for 43 parameters
In this case, the instability factor from reflection statistics seems to be reasonable because
the structure of a three-fold twin is fully solved and data are not expected to contain more
information.

107 Jana2006 Cookbook, version October 2017

108 Jana2006 Cookbook, version October 2017
Example 3.4: Ca2Fe2O5
HT structural phase transition, coexistence of two phases in single crystal with
strong overlaps.

H. Kruger, V. Kahlenberg, V. Petříček, F. Phillipp & W. Wertl, J. Solid State Chem., 182, (2009),
1515-1523.

LT phase – orthorhombic, Pnma, cell parameters ~ 5.4707, 14.9897, 5.6302 Å

In one layer of Fe tetrahedrons there is only one type orientation (e.g. Left)

HT phase (>710⁰C) – orthorhombic, modulated, Imma(00γ)s00, cell parameters ~ 5.4707,

14.9897, 5.6302 Å, q=(0,0,0.58). In one layer of Fe tetrahedrons there are two different
orientations (Left and Right) alternating with the period incommensurate to the basic cell.

109 Jana2006 Cookbook, version October 2017

Diffraction pattern: Strong 1st order, no 2nd order and some weak 3rd order satellites → crenel
like modulation with ∆ ≈ 1 2 .

Main features of the solution for the HT phase: The atoms Fe2 and O3 are occupying the
positions close to the Wyckoff position 4(e), point group mm2. The mirror m, with normal along
a axis, having half shift along the internal x4 axis, transforms the “left” oriented ribbon of Fe
tetrahedrons to the “right” ones and the occupational modulations of both atoms have a crenel
like shape.

During the first order phase transition both phases (LT and HT) are coexisting in a wide region
of temperatures and with x-ray diffraction they both contribute to the same diffraction pattern.
In this example we shall start from the known structures of both phases and we shall
demonstrate that the generalized twinning option in Jana2006 can also be applied to single
crystals composed from two or more phases.

As a starting point you have at disposal the HT structure refined against the data set having the
highest possible fraction of the HT phase from all measured data sets. In the first part of the
example it is demonstrated main features of the solution. In the second part the LT phase will
be read in from the supplied CIF file and then the final refinement will be performed.

Input files: HT_phase.m40, HT_phase.m50, HT_phase.m90, HT_phase.m95 and

LT_phase.cif.

1. Open the modulated phase

Start Jana2006
“File → Structure → Open” opens a file manager
Left pane: locate directory with input files
Right pane: double-click HT_Phase

2. Check systematically extinct reflections

“File → Reflection file → Create refinement reflection file”

There are 381 observed reflections, some of them relatively strong, which do not follow the
extinction rules for Imma(00γ)s00. This means that the reflections of the second
“P-centered” phase are really present.
110 Jana2006 Cookbook, version October 2017
3. Run refinement
[On the screen: basic window of Jana]
Right-click the icon of Refine.
[On the screen: refinement options]
Define 100 of cycles, Damping factor 1; OK
Choose “YES+START”
The refinement converges with R(obs) about 5.98%, 4.48% and 9.12%.

4. Check the refinement model

[On the screen: basic Jana window]
Start “Edit atoms”
[On the screen: list of atoms]
Double-click on the first atom Fe1
[On the screen: list of atoms]
Go though all atoms (using arrows of the top-left text box) and check the modulation
model.
All atoms except Fe2 and O3 have just one harmonic positional wave. Atoms Fe2 and O3
have one crenel function.
Write down coordinates of Fe2 and O3 and then go to “Edit” thumbnail, press the button
“Occupancy” and write down also the values of delta and x40
Fe2 (-0.052,0.25,0.568) delta=0.5 (fixed) x40~0.6404
O3 (-0.402,0.25,0.625) delta=0.5 (fixed) x40~0.6755
Escape the form;

5. Run dist
[On the screen: basic Jana window]
Right-click the icon of Dist.
[On the screen: distance commands]
Use the default parameters except for the listing form which should be changed to “With
symmetry code”.
OK; Start the program;
Open the listing and look for Fe2 coordination;
The atom Fe2 is coordinated by the atom O2, O2#s2, O3 and O3#s4c2t0,0,1.

6. Draw modulation curves in Contour for Fe2 and O3

[On the screen: basic Jana window]
Double-click the icon Contour; Press the button “New plot” if active;
Select “Fourier map” and “Draw maps as calculated”; OK
In thumbnail “Basic” select “F(obs) – Fourier” and in thumbnail “Scope” select “by a central
point”;
Center “0 0.25 0.568”; Section 1st → x1, 2nd → x4, 3rd → x2, 4th → x3;
x4 minimum → 0, maximum → 2, step 0.02
Scope → 3 0 0; Step → 0.02; OK
[On the screen: the following map]

111 Jana2006 Cookbook, version October 2017

Press “Atom edit” and add two atoms Fe2 and Fe2#s8. The second one is the position
generated by the eighth symmetry operation i.e. by -x1 x2 x3 x4+1/2;

[On the screen: the following map]

Repeat it for the center at -0.5 0.25 0.625 (a position close to O3);
Press the button “Atoms edit” to show the modulation curves for O3 and the symmetry
related atom O3#s8t-1,0,0. The translation part -1,0,0 returns the symmetry related atom

112 Jana2006 Cookbook, version October 2017

back to vicinity of O3.
[On the screen: the following map]

The crenel function for atoms Fe2 and O3 assures the alternation between the “Left” and
“Right” configurations of Fe tetrahedrons. During the refinement the “delta” values were
fixed to 0.5 for both atoms. But there were no restrictions applied to x40. Now we have to
check:

(a) if there is no overlap between two symmetry related positions of Fe2 and Fe2#s8 and O3
and O3#s8t-1,0,0.

The symmetry operation #8 has the following form: -x1, x2, x3, x4+1/2
for the atom Fe2 we have: t min = x40 − ∆ 2 − q ⋅ r t max = x40 + ∆ 2 − q ⋅ r and for the
symmetry related Fe2#s8: t min = x40 + 1 2 − ∆ 2 − q ⋅ r ′ t max = x40 + 1 2 + ∆ 2 − q ⋅ r ′
From the fact that the modulation vector is (0,0, γ ) and that the third component of the
symmetry related atoms remains the same for both positions, it follows that the t intervals
are fully separated with no overlaps.
The same conclusion is valid also for the atom O3.

(b) if the t intervals in which exist atom Fe2, O3 and O3#s4c2t0,0,1 are identical.

For Fe2 and O3 this means:

t min (Fe 2 ) = x40 (Fe 2 ) − ∆ (Fe 2 ) 2 − q ⋅ r (Fe 2 ) = t min (O 3) = x40 (O 3) − ∆ (O 3) 2 − q ⋅ r (O 3) =

taking into account the fact that ∆ (Fe 2 ) = ∆ (O 3) = 1 2 we have:

x40 (Fe 2 ) − q ⋅ r (Fe 2 ) = x40 (O 3) − q ⋅ r (O 3) ⇒ x40 (O 3) = x40 (Fe 2 ) − q ⋅ [r (Fe 2 ) − r (O 3)]

and finally x40 (O 3) = x40 (Fe 2 ) − γ ⋅ [z (Fe 2 ) − z (O 3)] (1)

113 Jana2006 Cookbook, version October 2017

 Numerically: x40 (O 3) = 0.6755 ≈ 0.6404 − 0.6101 ⋅ [0.568 − 0.625] = 0.6752 and
t min (O 3) = 0.6755 - 0.6101 ⋅ 0.625 = 0.0442

This means that the refined value was really close to the ideal value

For O3#s4c2t0,0,1 the symmetry operation s 4c 2t 0,0,1 is x1+1/2 -x2+1/2 -x3+3/2 -x4+1/2
modifies tmin to:

t min = x40 (O 3# s 4c 2t 0,0,1) − ∆ (O 3# s 4c 2t 0,0,1) 2 − q ⋅ r (O 3# s 4c 2t 0,0,1)

z (O 3# s 4c 2t 0,0,1) = 3 2 − z (O 3) = 0.825
Numerically:
x40 (O 3# s 4c 2t 0,0,1) = 1 2 − x40 (O 3) = −0.1755

This means that: t min = −0.1755 − 0.6101 ⋅ 0.875 − 0.25 = −0.959 and therefore we have to
modify the equation by adding one:

t min (Fe 2) = x40 (Fe 2) − ∆(Fe 2) 2 − q ⋅ r (Fe 2) =

x40 (O 3# s 4c 2t 0,0,1) − ∆(O 3# s 4c 2t 0,0,1) 2 − q ⋅ r (O 3# s 4c 2t 0,0,1) + 1 ⇒

x40 (O 3# s 4c 2t 0,0,1) = x40 (Fe 2 ) − q ⋅ [r (Fe 2 ) − r (O 3# s 4c 2t 0,0,1)] -1

and by applying the symmetry operation s 4c 2t 0,0,1 ~ x1+1/2 -x2+1/2 -x3+3/2 -x4+1/2:

1 2 − x40 (O 3) = x40 (Fe 2 ) − γ ⋅ [z (Fe 2 ) − 3 2 + z (O 3)] − 1 ⇒

x40 (O 3) = − x40 (Fe 2 ) + γ ⋅ [z (Fe 2 ) − 3 2 + z (O 3)] + 3 2 (2)

Combining (2) with the previous one we have for x40 (Fe 2 ) :

x40 (Fe 2 ) = 3 4 ⋅ (1 − γ ) + γ ⋅ z (Fe 2 ) ⇒ t 40 (Fe 2 ) = 3 4 ⋅ (1 − γ )

x40 (O 3) = 3 4 ⋅ (1 − γ ) + γ ⋅ z (O 3) ⇒ t 40 (O 3) = 3 4 ⋅ (1 − γ )

Numerically: x40 (O 3) = 0.6755 ≈ 0.75 ⋅ 0.3899 + 0.6101 ⋅ 0.625 = 0.6737

This means that both t40 parameters can be fixed to the same constant value as Fe2.

7. Run restricted refinement

[On the screen: basic window of Jana]
Right-click the icon of Refine.
[On the screen: refinement options]
114 Jana2006 Cookbook, version October 2017
 Select thumbnail “Various” and press the button “Equations”:
Define equations:
t40[Fe2]=0.75-0.75*cq3
t40[O3]=0.75-0.75*cq3
OK; Start Refinement;
The refinement converges with R(obs) about 5.99%, 4.49% and 9.14% which means that it
had almost no effect.
Open the listing to check the applied equation:

Equations defined by user:

x40[Fe2]=0.2924+0.6101*z[Fe2]
x40[O3]=0.2924+0.6101*z[O3]

8. Continue with a new structure name

[On the screen: basic window of Jana]
Start Jana2006
“File → Structure → Save as” opens a file manager
Right pane: Write a new structure name e.g. “Both_Phases”
[On the screen: “Continue with the new structure?”]
Yes;

9. Adding of the LT phase

[On the screen: basic window of Jana]
“Tools → Phases → New phase”;
Select “Import from CIF format”; Press “Browse” and select the file LT_Phase.cif;
OK

10. Define the twinning matrices

[On the screen: the form for defining of twinning matrices]
The “twinning” matrices define relationship between the phases and the reference cell
parameters. Unlike real twinning they may change the unit cell. In our case both phases are
defined in the same cell.
Both matrices are unit ones. The first is to be applied to the first phase and the second to
the second phase.
OK; follow the wizard for creating the refinement reflection file
Systematic extinctions are now much better fulfilled:

115 Jana2006 Cookbook, version October 2017

11. Run refinement
[On the screen: basic window of Jana]
Right-click the icon of Refine.
[On the screen: refinement options]
Change the damping factor to 0.5.
OK; Save and run the refinement;
The refinement converges with R(obs) about 4.99%, 4.09% and 7.37%.

12. Refine anisotropic ADP of atoms from the second phase

[On the screen: basic window of Jana]
Switch the phase to the Phase#2 (thumbnails in the lower-left corner)
Using “Edit atoms”: select all atoms and change ADP to harmonic;
OK;

13. Run refinement

[On the screen: basic window of Jana]
Double-click the icon of Refine.
The refinement converges with R(obs) about 4.60%, 3.78% and 6.75%.

14. Check the phase fraction factor

[On the screen: basic window of Jana]
“Parameters → Twin fractions”
The fraction of the second domain was refined to 0.276

116 Jana2006 Cookbook, version October 2017

Example 4.1: PtCu
Disordered structure.
Revised 26 September 2017

[Cu(CH3-NH-CH2-CH2-NH2)2] [Pt(CN)4]

Polyhedron 2011, 30, 269-278.

Single crystal data measured with Oxford Diffraction four-circle diffractometer

Input files: PtCu.hkl, PtCu_red.sum
Frame scaling, absorption correction: done with software of diffractometer

1. Creating new jobname

Start Jana2006
“File → Structure → New” opens a file manager
Left pane: locate directory with input files
Right pane: double-click PtCu

2. Import Wizard
Select “Known diffractometer formats”; NEXT
Select “Oxford Diffraction-CCD”; NEXT
[On the screen: “Define basic input file”]
Use "input from sum file"; OK
[On the screen: “Your sum file consists of 6 appended data reduction runs”]
OK;
[On the screen: “Complete/correct experimental parameters”]
Leave all settings unchanged; NEXT
[On the screen: “Define the reference cell/split by twinning”]
Leave all settings unchanged; NEXT
For absorption correction select “None or done before importing”; NEXT
FINISH

3. Data Repository
The line in data repository window corresponds to one data set.
OK; YES to accept the data set

4. Symmetry Wizard
[On the screen: “Tolerances for crystal system recognition”]
Leave all settings default; NEXT
[On the screen: Select Laue symmetry]
Select Tetragonal Laue symmetry 4/m; NEXT
Select primitive unit cell; NEXT
[On the screen: Select space group]
Using the criteria described in the first example Select P42/m; NEXT
Select "Accept the space group transformed into the original cell"
Clear “Accept twinning matrices induced by the space group test”
117 Jana2006 Cookbook, version October 2017
 The highest Laue symmetry consistent with the unit cell parameters is 4/mmm while the
selected Lause symmetry is lower – 4/m. This opens possibility for twinning, but in this case
twinning does not occur (as one would find during refinement of twin volume fractions).
FINISH
The original cell is used because the suggested transformation does not change the unit cell
and the space group symbol
YES to create refinement reflection file M90

5. Creating refinement reflection file

NEXT to confirm threshold 3sigma
[On the screen: settings for merging of reflections]
Leave all settings default; NEXT
Accept the data block; FINISH

6. Structure Solution Wizard

[On the screen: window of Structure solution]
In “Formula” textbox type chemical composition C10 H20 Cu Pt N6, two formula units
Select "Use Superflip"
Check “Repeat Superflip”, number of runs 10
Leave other settings default; “Run Solution”
[On the screen: listing of Superflip]
Superflip converges rapidly with R~14% and very good (i.e. low) fit for the symmetry
elements of 42/m
The suggested centering vector (0,0,1/2) is in contradiction with found systematic absences,
it cannot be used. It is a pseudosymmetry caused by positions of platinum (heavy atom).
Close listing of Superflip
Press “Accept last solution” to leave the Structure solution wizard

7. Completing the structure model

Plot the structure found by Superflip
There should be a chain Pt – C – N – Cu visible in the plot.

Open Refinement options, change number if refinement cycles to 100

Run Refine
Refinement converges with R value about 12%
Start “Edit atoms” and define harmonic ADP for Pt1 and Cu1
Delete rejected atoms (in case the button "Select rejected" is active)
Run Refine
Refinement converges with R value about 9%
Open Fourier options (right button on Fourier icon) and select “F(obs)-F(calc) - difference
Fourier”, OK, YES+Start to execute Fourier calculation
YES to start the procedure for including of the new atoms
118 Jana2006 Cookbook, version October 2017
The same tool could be started separately by “Parameters → Atoms → New”
[On the screen: Inserting/replacing of atoms]
For “Show distances up to” type 4 Å

OK
[On the screen: list of peaks]

In the left panel, Max1 is focused. Its integrated charge is displayed at the bottom. In the
right panel, we can see distances of symmetry equivalent positions of Max1 to already
existing atoms of the structure model, up to the previously defined limit 4Å. We can expect
distances ~2 Å for Pt-C, ~1.5 Å for C-C, ~1 Å for C-N of cyano group and ~3 Å for Pt-N.
Therefore, the maximum Max1 is probably nitrogen.
In the left panel, focus Max1 (or other peak if in your case Max1 has no good distance)
In the right panel, focus its reasonable position having good distance to some existing atom
Press “Include selected peak”
[On the screen: Complete information for the new atom]
For “Name of atom” type C* (even for nitrogen); TAB; OK
We will intentionally add all atoms like carbon, it will help later on to analyse disorder.
The chemical type was assigned to carbon, the name will be C+sequence number.
Repeat the procedure for other maxima. Skip distances below 1 Å as well as maxima with
too small charge or strange distances. At this point, all atoms should be added as carbon,
we will distinguish between C and N later on.
FINISH; YES to include the new atoms
Finally the structure model should contain 9 symmetry-independent atoms. In case some
improper maxima have been added, they will be probably discarded during refinement
based on large isotropic ADP.
Run Refine
Refinement converges with R value about 4%
Start “Edit atoms” and check if the button “Select rejected” is enabled.
If “Select rejected” is enabled some atoms have zero occupancy. It was set by Refine because
their Uiso was refined above the limit defined in the Refinement options (default limit is 0.2).
Delete rejected atoms.

119 Jana2006 Cookbook, version October 2017

 Check with “Select all” that the structure contains 9 atoms
If the structure model does not contain all 9 atoms, repeat calculation of difference Fourier
Leave “Edit atoms”
Plot structure in Diamond

Note names of atoms (C1, C5 in our example) which should be nitrogen of cyano groups
Leave Diamond and change chemical type and names of C1, C5
Run Refine

8. Solution of disorder
Open “Edit atoms” and define harmonic ADP for all atoms
Double-click is necessary in the locked yellow-indicated options
Run Refine
Refinement converges with R value ~3% for 60 parameters
Plot the disordered part of the structure: press ; click on Pt; Ctrl-M selects the Pt
“molecule”; Delete deletes it – remaining atoms are coodinated to Cu

Using “Action → Rename selected atoms manually”, rename atoms according to this plot
This is not necessary but will help to follow easily the cookbook
Start “File → Structure → Save As” and create a backup copy, for instance PtCu_completed
Open “Edit atoms”, select C6
“Action” → “Split atomic position”
[On the screen: split atomic position window]
Leave all settings default; Apply
[On the screen: list of atoms in “Edit atoms”]
Select C6, go to “Edit/define atoms”, switch to the page “Edit”, activate refinement of
occupancy (the parameter ai)
120 Jana2006 Cookbook, version October 2017
Close "Edit atoms", YES to save changes
The next steps show automatically created restriction for C6, C6’
Open Refinement options, go to the page “Various”, press “Restrictions”
Focus the line “restric C6 2 C6’
Press Load
In the lower part of the window there are listed properties of the restriction. The atom C6
was split to two positions, C6 and C6’. Both atoms will have the same ADP and their
occupancy will be complementary.
Leave Refinement options
Run Refine
Refinement converges with R=~2% for 59 parameters
Plot the same part of the structure

Open “Edit atoms”

Using “Action → Rename selected atoms manually”, rename atoms according to this plot
This is not necessary but will help to follow easily the cookbook
Enable refinement of occupancy for C2 and C4
Run Refine
Refinement converges with R=1.8% for 61 parameters
Note refined occupancies
Plot the same part of the structure and compare occupancies of carbons
(Diamond shows occupancies when hovering the mouse on the atom)

121 Jana2006 Cookbook, version October 2017

We are looking for two symmetry-related half-occupied methyl-ethylenediamine chains.
Atoms with occupancy above 0.5 will be half-occupied atoms of nitrogen. In positions with
occupancy above 1 atoms of two disordered chains coincide. Occupancy of C2 and C4
positions also reflects the missing hydrogen atoms while occupancy C6, C6’ keeps the sum
1.0. Based on these considerations, we can speculate that the methyl-ethylenediamine
chains will be oriented like in the next figure:

Open “Edit atoms” and change chemical type of C4 and C6 to nitrogen.

Using “Action → Rename selected atoms manually”, rename C4 to N4 and C6 to N6 and C6’
to C6.
The atom names in the restrictions (page Various of the refinement commands) are
renamed automatically
Run Refine
Open “Edit atoms”
Change occupancy of C2 to 1
Split the position of C2 (using the same procedure like previously for C6)
Enable refinement of occupancy for C2
The reason for changing the occupancy to 1 is that the generated restriction will keep the
sum of actual occupancies.
122 Jana2006 Cookbook, version October 2017
 Run Refine
A slight increase of R value is due to change of ADP to isotropic during splitting.
Open “Edit atoms” and change ADP of C6 and N6 from isotropic to harmonic
Open Refinement options, go to the page "Various", press "Restrictions"
Load the restriction C6-N6 and uncheck “ADP parameters”; press “Rewrite”
The atoms C6 and N6 are so far from each other that their ADP can be refined independently
OK; OK; YES+Start to save changes and run refinement
Refinement converges with R=1.7% for 70 parameters
Plot the structure
Atom N4 represents through its symmetry equivalent positions the terminal nitrogen of both
methyl-ethylenediamine groups. From this follows that the chains are symmetry-equivalent
and must have the same occupation 0.5
Open “Edit atoms” and select all atoms of both methyl-ethylenediamine groups
(N4,C2,C2’,C6,N6 in our example)
Rewrite their occupancy to 0.5 and disable its refinement
Double-click will be necessary for the locked yellow-indicated fields
Run Refine
Refinement converges with R=1.7% for 67 parameters
Create backup copy, for instance PtCu_beforeH
Plot the result

9. Adding hydrogen atoms

For adding hydrogen atoms the automatic procedure cannot be used because of the
disorder. In the following plot, we can find how many hydrogen atoms we need for carbons
and nitrogens. Warning: In your structure model, the distances may slightly differ because of
missing hydrogen atoms. It may also happen that C2 and C2’ are swapped (in such case it is
better to rename them).

C2 belongs to methyl
C2’ and C6 belong to CH2
N4 belongs to NH2
N6 belongs to NH
123 Jana2006 Cookbook, version October 2017
Open “Edit atoms”
Select C2, C2’, C6, N4 and N6
Start “Action → Adding of hydrogen atoms”
Clear “Try automatic run”; OK
[On the screen: Adding hydrogen atoms for C2]
C2 belongs to methyl, therefore it has one neighbor and three hydrogen atoms
Select “Tetrahedral”
Change number of neighbors to 1
Click in the textbox for the neighbor: this activates the button “Select neighbors” in the
lower right corner
Press “Select neighbors”
Select N6; OK
In the textboxes for hydrogens type H1C2, H2C2 and H3C2
Press “Apply → Go to next”
[On the screen: Adding hydrogen atoms for C2’]
C2’ belongs to CH2, therefore it has two neighbors and two hydrogen atoms
Select “Tetrahedral”
Change number of neighbors to 2
Click in the textbox of the first neighbor; button “Select neighbors” will be activated
Press “Select neighbors”
Select N4 and C6; OK
In the textboxes for hydrogens type H1C2’ and H2C2’
Press “Apply → Go to next”
Similarly, continue for C6. For C6 two atoms N6 will be shown as neighbors but the one with
too short distance belongs to the second chain
Similarly, continue for N4 belonging to NH2. Here one of the neighbors will be copper.
Similarly, continue for N6 belonging to NH. Geometry for N6 is trigonal, not tetrahedral.
Leave “Edit atoms”
Check in Diamond geometry of hydrogens (needs manually deleting of unwanted distances
using Table of created bonds)
Run Refine
Refinement converges with R value 1.6% for 67 parameters

At this structure model, hydrogen atoms attached to nitrogen are strictly kept in idealized
geometry. This is correct for carbon but not for nitrogen. In the next step, we will redefine
constraints (keep commands) to restraints.
Start refinement commands; select the page Various; open “Keep commands”
Disable the command “keep hydro tetrahed N4”
Disable the command “keep hydro triang N6”
The commands “keep ADP riding N4” and “keep ADP riding N6” will be retained
OK; open “distance restraints”
Select “Fix to the specified value” and for the “Value” type 0.87
In the textbox “Atoms”, type: N4 H1N4; N4 H2N4; N6 H1N6
Press “Add”; OK; OK; “YES+START”
In case of bad convergence, change the dampling factor to 0.5.
Refinement converges with R value ~1.5% for 76 parameters

124 Jana2006 Cookbook, version October 2017

This plot in Diamond requires a lot of manual work

125 Jana2006 Cookbook, version October 2017

126 Jana2006 Cookbook, version October 2017
Example 4.2: NaNiFe
Structure with mixed sites.
Revised: 08 March 2012

Expected formula: NaNiFe2P2O12

Single crystal data measured with Oxford Diffraction four-circle diffractometer

Input files: NaNiFe.hkl, NaNiFe_red.sum
Frame scaling, absorption correction: done with software of diffractometer

1. Creating new jobname

Start Jana2006
“File → Structure → New” opens a file manager
Left pane: locate directory with input files
Right pane: double-click NaNiFe

2. Import Wizard
Select “Known diffractometer formats”; NEXT
Select “Oxford Diffraction → CCD”; NEXT
Select “Input from cif_od file”; OK
Leave all settings unchanged; NEXT
Leave all settings unchanged; NEXT
7999 reflections has been imported
For absorption correction select “None or done before importing”; NEXT
FINISH
[On the screen: Data repository]
OK; YES to accept the data set

3. Symmetry Wizard, merging of reflections

Go through the steps of the Symmetry wizard, select Laue symmetry 2/m, C centering and
space group C2/c
Accept the space group transformed into the original cell; FINISH
Go through the steps of the Merging wizard leaving all settings default
Rint will be 2.57%

4. Structure Solution Wizard

[On the screen: window of Structure solution]
In “Formula” textbox type chemical composition Na Ni Fe P O, 4 formula units
Select “Use Superflip”
Select “Use a specific random seed” and for the random seed define 111
Fixing the random seed to 111 guarantees for most processors that the Superflip returns
exactly the same results like described in this cookbook. For normal work random seed
should not be fixed
Leave other settings default; OK
[On the screen: listing of Superflip]
Superflip converges rapidly with R~18% and very good (i.e. low) agreement factors for the
127 Jana2006 Cookbook, version October 2017
 symmetry C2/m
Close listing of Superflip and accept the results

5. Checking the structure model

Start EditM50, page Composition, press “Formula from M40”
It shows the formula “Na Ni3 Fe0 P3 O13” (for Z=4). However, the expected disphosphate
unit is P2O12 with negative charge -9. The positive charge NaNi3 is only 7. Why do we have
13 oxygens? Where is Iron?
Leave to the basic window of Jana2006
Start refinement
Refinement converges with R=5.7%.
Using “Edit atoms” change all ADP to harmonic and repeat the refinement
Refinement converges with R=4.4%
Open Fourier options and select Difference Fourier
“Yes+Start” to save changes and start Fourier
YES to include new atoms
The list of Fourier maxima does not display any useful maximum. The structure is complete.
Return to the basic window of Jana2006 without adding new atoms
Open refinement listing
Go To → list of serious warnings
The listing says that ADP of Na1 and O1 are not positive definite
Return to the basic window of Jana2006

6. Calculation of distances and bond valences

Open Distances options (right-click the icon Dist)
Uncheck “d(max) derived from…”
For “d(max)” type 3.5
Press the button “Define coefficients for bond valences”
Select “1st atom type” Na; “2nd atom type” O
Press the button “From file” and select the first line for bond valences coefficients
The list shows the coefficients like published in various sources. We are interested only in
differences in coordinations and not in absolute values of bond valences sums. Therefore we
shall use the first line for all cases.
Press Add
Do the same for the pairs Ni-O, P-O and Fe-O

OK; OK; Yes+Start

Yes to open listing
Go To → List of distances
128 Jana2006 Cookbook, version October 2017
 Go To → List of bond valences
Summary (“BVS=bond valence sum”):
Ni1 is coordinated by six oxygens, Ni-O between 2.10 and 2.13 Å, BVS 1.75
Ni2 is coordinated by six oxygens, Ni-O between 1.95 and 2.12 Å, BVS 2.21
P1 and P2 are quite normal
Na1 is coordinated by eight oxygens, Na-O between 2.27 and 2.83 Å, BVS 1.43
O1 is coordinated by eight oxygens, O1-O between 2.38 and 2.87 Å
ADP of Na1 and O1 are negative
In plotting software we can get following picture with O1 in the void:

7. Refinement of occupancy Na1, O1

Using “Edit atoms” select that occupancy of Na1 and O1 will be refined

Start refinement
Refinement converges with R=3.4%
Occupancy of Na1 refines to 0.62
Occupancy of O1 refines to 0.69
From refinement listing we can see there are no warnings about negative ADP
Na1 must be mixed with Ni or Fe because the refined occupancy exceeds the site occupancy.
The BVS larger than 1 support this idea.
For oxygen 0.69*2*8=11; 2 is the site multiplicity, 8 is number of electrons. The result is not
too far of 10 which is sum of electrons for water. Therefore, O1 could be water.
Using “Edit atoms” change chemical type of O1 to Na without renaming
Run refinement
Calculate distances
R value is the same. Occupancy of O1 (now chemical type Na) refines to 0.46. BVS of O1 is

129 Jana2006 Cookbook, version October 2017

 1.0. Therefore O1 could be also sodium, but partially occupied because there is no lighter
element available for mixing.
In the next part we shall consider O1 to be oxygen of a water molecule.

8. Mixed site Na1-Ni3

Using “Edit Atoms” change chemical type of O1 back to oxygen
Run refinement
Using “File → Structure → Save As” create a backup copy of the structure but continue
with the original one
Jana does not have a tool for creating mixed sites. It should be done in M40 by text editor.
Using “Edit/view → Edit M40 file” open M40
Double the two lines belonging to Na1
For the new atom change “Na1” to “Ni3”
In the header change “12” to “13”
Save and close M40
Jana checks the syntax of the changed M40. No warnings mean correct syntax.

Start “Atoms Edit”

Change chemical type of Ni3 to nickel
Change site occupancy of Na1 and Ni3 to 0.25 and activate their refinement keys
We need sum of occupancies equal to 0.5
Close “Edit Atoms” and open Refinement options
Switch to page “Various” and press “Restrictions”
For “Atoms/Molecules” type “Na1 Ni3”
Select “Coordinates”, “ADP parameters” and “keep overall sum”; press “Add”
It creates the restriction “restrict Na1 1 Ni3” with the following meaning: both atoms will
have the same coordinates, the same ADP and constant sum of site occupancies
OK; OK, “Yes+Start” to run refinement
Refinement converges with R value still around 3.4%
Site occupancies refine to 0.43 for Na1 and 0.07 for Ni3
Similar results could be achieved with iron so at this point the chemical type of the metal is
either nickel or iron

130 Jana2006 Cookbook, version October 2017

9. Mixed sites Ni1-Fe1 and Ni2-Fe2
Using “Edit atoms” activate refinement of site occupancies for Ni1, Ni2
Run refinement
Refinement converges with R=3.1% (improvement)
Occupancy of Ni1 refines to 0.47 (from 0.5)
Occupancy of Ni2 refines to 0.94 (from 1.0)
In refinement listing – “changes overview” – we can check that s.u. of occupancy parameters
is small.
Maybe Ni1 and Ni2 are mixed with iron
Using “Edit atoms” change occupancy of Ni1 to 0.25 and occupancy of Ni2 to 0.5
With these numbers the doubled position will have the sum of occupancies equal to the site
occupancy
Using “Edit/view → Edit M40 file” double the two lines of Ni1 and also double the two lines
of Ni2
Change the second “Ni1” to “Fe1” and the second “Ni2” to “Fe2”
Increase number of atoms in the header by 2; close the text editor
Using “Edit atoms” change chemical type of Fe1 and Fe2 to iron; make sure that site
occupancy refinement keys are activated for Ni1,Fe1,Ni2,Fe2
Open Refinement options, page Various
Define restriction for the pair Ni1-Fe1 and for the pair Ni2-Fe2 (by the same way like
previously for Na1-Ni3, keeping the overall sum)
Run refinement
Refinement converges with R value 3.1%
Occupancy of the pair Ni1-Fe1 refines to 0.24, 0.26
Occupancy of the pair Ni2-Fe2 refines to 0.36, 0.64

10. Calculation of the charge

Start “EditM50”, page “Composition”, and press “Formula from M40”
It prints Na0.878 Ni1.309 Fe1.812 P3 O13.331
O13.331 is O12 when we subtract the “water”
Assuming that the valence of iron is three we can calculate: 1.81*3+1.31*2+0.88=8.93. This
is very close to the required charge 9.

By this way we have created one possible model. Unfortunately there is a lot of another
possibilities which should be checked, for instance:

The water could be partially occupied sodium (it slightly decreases R value to 3.07%) and the
valence of the iron mixed with Ni could be 2+. The composition would be Na1.769 Ni1.27
Fe1.844 P3 O12.
1.844*3+1.27*2+1.769=9.84, i.e. exceed of positive charge

For the above case we could also consider mixed valence for Fe. To get easily the formula we
can temporarily change chemical type of Fe1 to something else, for instance aluminium. But
first it must be added to the end of the formula (to become available in “Edit atoms”).
EditM50 then calculates the formula Na1.769 Ni1.269 Fe1.324 P3 O12 Al0.52.
For Al representing trivalent iron the charge would be 0.52*3*1.32*2+1.27*2+1.769=8.51,
i.e. lack of positive charge
For Al representing divalent iron the charge would be 0.52*2*1.32*3+1.27*2+1.769=9.31,
131 Jana2006 Cookbook, version October 2017
i.e. exceed of positive charge

For the above case with lack of positive charge (sum 8.51) we can also consider mixing of
Na1 not with nickel but with trivalent iron. It yields: Na1.748 Ni1.153 Fe1.325 P3 O12
Al0.656 with trivalent iron represented by aluminium. The charge is
0.66*3+1.33*2+1.15*2+1.75=8.69, i.e. still lack of the positive charge

The Ni1 site is maybe not mixed (BVS is smaller than for Ni2 and mixing has small impact on
R). But iron alone in this site increases R value. It can only be nickel alone yielding the
formula: Na1.716 Ni1.713 Fe1.437 P3 O12. The charge would be 9.46 (trivalent iron), 8.02
(divalent iron), 8.2 or 9.3 (for mixed valence of iron).

It seems that another model with proper charge is not easy to find (maybe it will be
discovered during the workshop). Good proof would be to find hydrogen atoms of water in
the difference Fourier map but they are not there. Or we could find hydrogen in the P2O12
group which would allow the more logical solution with positive charge 8.02, but it is also
not present in difference map.

Conclusions:

The formula is either

Na0.878 Ni1.309 Fe1.812 P3 O12 . H2O with trivalent iron

or

Na1.716 Ni1.713 Fe1.437 HP3 O11 OH with divalent iron

The latter yields slightly better R value and corresponds better to found bond valences sums.
However, hydrogens cannot be determined from the data.

(Fast microprobe analysis yields Na1.56 Ni1.69 Fe0.78 P3 O6 which is insufficient precision
for making a decision)

132 Jana2006 Cookbook, version October 2017

Example 5.1: YPO
Simple modulated structure with crenel
Revised 14 December 2015

Y(PO3)3
Single crystal data measured with Oxford Diffraction four-circle diffractometer
Input files: YPO.hkl, YPO_red.sum
Frame scaling, absorption correction: done

1. Creating new jobname

Start Jana2006
“File → Structure → New” opens a file manager
Left pane: locate directory with input files
Right pane: double-click YPO

2. Import Wizard
Select “Known diffractometer formats”; NEXT
Select “Oxford Diffraction → CCD”; NEXT
[On the screen: “Define basic input file:”]
Select “Input from sum file”; file name “ypo_red.sum”; OK
[On the screen: “Warning about appended data reduction runs”]
OK
[On the screen: “Complete/Correct experimental parameters”]
Leave all settings unchanged; NEXT
[On the screen: “Define the reference cell/split by twinning”]
Leave all settings unchanged; NEXT
For absorption correction select “None or done before importing”; NEXT
FINISH

3. Data Repository
The line in data repository window corresponds to one data set.
OK; YES to accept the data set
[On the screen: “Information about starting symmetry wizard”]
NEXT

4. Symmetry Wizard
[On the screen: “Tolerances for crystal system recognition”.]
Symmetry wizard can be started separately by “File → Reflection file → Make space group
test”.
Leave all settings default; NEXT
[On the screen: Select supercell]
Select "Continue with the basic cell"; NEXT
[On the screen: Select Laue symmetry]
Select Monoclinic Laue symmetry; NEXT
[On the screen: Select cell centering]

133 Jana2006 Cookbook, version October 2017

 Select I centered unit cell; NEXT
[On the screen: Select space group]
Select C2/c; NEXT
[On the screen: Select superspace group]
C2/c(0b0)s0
[On the screen: Final step of the superspace test]
Accept the superspace group in the standard setting: C2/c(0b0)s0; FINISH

5. Creating refinement reflection file

We have finished creation of the file M95 (reflections converted from diffractometer format
to Jana format) and M50 (cell parameters, symmetry). At this step we shall create the file
M90 containing data set merged by symmetry, with extinct reflections omitted.
NEXT to confirm threshold 3sigma
[On the screen: settings for merging of reflections]
Leave all settings default; NEXT
Accept the data block and calculate coverage; FINISH
[On the screen: Information that structure solution will start]

6. Structure Solution Wizard

[On the screen: window of Structure solution]
In “Formula” textbox type list of chemical elements for Y(PO3)3: Y P O
Select “use Superflip”; “Peaks from Jana2006”
Select "Use a specific random seed" and type value 5555
“Run solution”
Superflip converges (after noise suppression) with R value 21%. It confirms the symmetry
C2/c(0b0)s0.
Fixing the random seed to 5555 guarantees for most processors that the Superflip returns
the same results like described in this cookbook. Otherwise, each run of Superflip would
return slightly different electron density. For normal work, random seed should be automatic
because the convergence of charge flipping is not guaranteed for every value of random
seed.
Close the listing and press “Accept last solution” to leave the structure solution wizard

7. Verification of the structure solution

Start “Edit/View → editing of M40 file”. It shows the structure parameter file
Start “Edit atoms”. It shows the same in a graphical interface; parameters of atoms can be
accessed by double clicking on atom or by “Action → Edit/Define atoms”;
OK
Start EditM50, go to page Composition, change “Formula units” to 4
Press “Formula from M40”
[On the screen: Information about formula]
With previously used “randomseed 5555” the formula should be Y P3 O8 (for Z=4).
OK; YES for updating the formula
Modify the formula to Y P3 O9; OK, YES to rewrite M50 file
Correct numbers of elements helps to find correct number of atoms in Fourier maps. It is also
important for the final CIF generation.
Right-click the icon of Fourier and select thumbnail “Basic”. Choose “Map type”

134 Jana2006 Cookbook, version October 2017

“F(obs) - Fourier” and in the thumbnail “Scope” select the following:

OK; YES+START to save Fourier options and execute Fourier

Fourier reads as an input M80 with structure factors created previously by Superflip. M80 is
re-created after regular end of structure refinement. In this case refinement did not run so
M80 originates from Superflip.
Run Contour; start “New plot” – “Use old maps”+”Draw maps as calculated”.
Then the modulation density will appear. With “Atoms edit” define the Y1 atom to be
visualized.

The plot illustrates that the electron density from Superflip contains reliable information
about position modulation. The modulation curve results by interpretation of the electron
density map by Jana.
Quit Contour

135 Jana2006 Cookbook, version October 2017

8. Origin shift to the position of Y
The atom Y should be localized at the inversion center in one of the following Wyckoff
positions:
4d (1/4,1/4,1/2); (3/4,1/4,0) + (1/2,1/2,0)
4c (1/4,1/4,0); (3/4,1/4,1/2) + (1/2,1/2,0)
4b (0,1/2,0); (0,1/2,1/2) + (1/2,1/2,0)
4a (0,0,0); (0,0,1/2) + (1/2,1/2,0)
In the next step we shall shift the origin to the position of the Y atom. This step is not
necessary for the structure solution and refinement but it assures that the coordinates and
plots will look as described in this cookbook
Start “Tools → Transformations → Origin shift”
[On the screen: Move the origin to:]
Select Atom Y1 and press OK.

9. Verification of modulated functions found by Superflip

Right-click the icon of Refine and select thumbnail “Various”. Use the button “Fixed
commands”, select “Modulation parameters” and type * to the Atoms/parameters textbox.
This fixes modulation parameters for all atoms.
Press the button “Add”; OK; OK
Start Refinement.
The refinement should finish with R(obs) factors about 13%, 11%, 15% and 22% for all, main,
1st order satellite and 2nd order satellite reflections. This means that we have reasonably

good starting point to continue the solution.

The screen output of Refine is closed after a delay. It is saved in refinement listing
(“Edit/View → View of Refine → Go to → R-factors overview”)

10. Refinement of the modulated structure

Right-click the icon of Refine and select thumbnail “Various”. Use the button “Fixed
commands” and disable or delete the previously used “Fixed command”
Run refinement.
You should get R values about 12%, 10%, 14% and 21%.
Start “Edit atoms”
Press “Select all” and “Action → Edit/define”
Choose “harmonic” for ADP parameters.
OK; OK; Yes to rewrite changes
Start Refinement.
Refinement converges with R value about 11%, 8%, 12% and 20% for 94 parameters.

11. Localization of the missing oxygen atom

Right-click the icon of Fourier and select thumbnail “Basic” – select “Map type”
“F(obs)-F(calc) – difference Fourier”
In the thumbnail “Scope” select: “Automatically” and “Use default map orientation”
OK
Run Fourier
The modulation curve of this additional atom is rather complicated and therefore we have
to look at the density maps to see more details. First, we should check if the peak,
representing possible position of the new oxygen atom, has been really localized in the
difference Fourier map.

136 Jana2006 Cookbook, version October 2017

[On the screen: Do you want to start the procedure for including of new atoms?]
Answer yes.
[On the screen: Inserting/replacing of atoms – setting of distance limits]
Use default 3 Å; OK
(The dialogue for adding maxima is also available through “Parameters → Atoms → New”)
The dominating maximum (with the largest charge) should be close to the special position at
(-1/4,1/4,0) or to some of its equivalents, with reasonable distances to the atom P1 (about
1.5 Å).
Note the coordinates of the dominating maximum in equivalent position close to P1
For the current structure model (random seed 5555) it should be -0.245, 0.254, -0.503
Round the coordinates to the closest special position and note them
For the current structure model (random seed 5555) it should be: -0.25, 0.25, -0.5
Leave the form without adding the new atom.
Adding new atoms would add oxygen to a general position because its special position is not
properly recognized.
Use Contour to see difference Fourier sections through the point (-1/4, 1/4, -1/2).
Start Contour, press “New plot”
Select “Calculate new ones” and “Draw maps as calculated”; OK
In Basic thumbnail select “F(obs)-F(calc) – difference Fourier” and in the thumbnail “Scope”
define the vicinity of the point (-1/4, 1/4, -1/2):

OK starts Fourier calculation

The first section is immediately plotted:

137 Jana2006 Cookbook, version October 2017

The resulting map must be summed over x2, x3 direction to see the correct modulation curve
Press “Sum ON”; select “x2” and “x3” leaving the intervals defaults; OK
The resulting map shows the shape of modulation:

Analogously (starting with button “New plot” followed by “Calculate new ones”), calculate
and plot the section x2-x4 and the section x3-x4:

138 Jana2006 Cookbook, version October 2017

 The modulation has a discontinuity at the point with x4=0. This is clearly visible from the
map showing evolution of the x2 atomic coordinate (the x2-x4 section).

As the difference Fourier is more sensitive to light oxygen atom we should save the
temporary result for a final check of the refined modulation of that oxygen atom.

Close “Contour”
“File → Structure → Save as” and use a new name for example “YPO-for-maps” but do not
continue with the new structure.

12. Inserting the missing oxygen atom

From the previous we know that the modulation function of the new atom at (-1/4,
1/4, -1/2) has a discontinuity and therefore a crenel function will be used. In the x2-x4
difference section we found that the length (delta) of crenel is one and its center (x40) is
located at 0.

“Parameters → Atoms → New”, select “Coordinates from keyboard”; OK

Type the coordinates -1/4 1/4 -1/2, press “Use new coordinates”
Select the symmetrical position making the shortest distance to P1 and press “Include
selected peaks”.
[On the screen: Complete information for the new atom]
For “Name of atom” type “O*” followed by TAB
This selects the chemical type oxygen and the label with the first available sequence
number.
OK; Finish; YES to include new atom

Now we have to define the crenel function

139 Jana2006 Cookbook, version October 2017

Run “Edit atoms” and select the new atom (probably O5).
“Action → Edit/define” check “crenel” and add also one position modulation wave in the
textbox “Position”

Switch to the page “Edit” and press the button “Occupancy”, which is close to “Edit
modulation parameters”
define “1” for “delta” and “0” for “x40”; OK
This is width and centre of the crenel function
Go to the page "Define"
Select "Legendre polynomials in crenel interval"
OK; OK; Yes to rewrite

Draw Contour section x4-x2 for (-1/4 1/4 -1/2) - like previously
Press "Atoms edit" and add O5; OK

140 Jana2006 Cookbook, version October 2017

 The Contour plot shows the initial modulation function for O5

13. Refinement of the complete structure

Right click on the Refine icon, switch to “Basic”
Define there number of cycles 100 and damping factor 0.2.
Damping factor multiplies the parameters changes
Start refinement

→ R about 9%, 6%, 10%, 20%

Draw Contour section (without Fourier calculation!) x4-x2 for (1/4 -1/4 1/2) - like previously
Press "Atoms edit" and add O5; OK
The Contour plot shows the crenel modulation function for O5 combined with one
modulation wave

Run “Edit atoms”, select all atoms, press “Action” and select “Edit/Define atoms”
Modify all ADP parameters to harmonic ones by double-clicking the corresponding radio
button
Yellow indicates locked objects. Locking occurs when the corresponding parameters are not
the same (or of the same type) for the selected atoms. In our case ADP is harmonic for most
of atoms but isotropic for the newly added O5, this is the reason for locking of radio buttons.
The text box for site occupancy is also locked because some atoms are in special positions.

141 Jana2006 Cookbook, version October 2017

 A locked object can be unlocked by double-click.
Add another one positional modulation wave for all atoms using the textbox “Position”.
Save changes and start refinement

→ R about 5.6%, 3.4%, 6%, 18%

Run “Edit atoms”, select all atoms and (using the button “ADP 2nd”) define two ADP
modulation waves for all atoms.
Start refinement

→ R about 3.4%, 2.5%, 3.5%, 9.0% for 297 refined parameters

14. Check modulation curves with maps for O5

In this part we will check whether the refined modulation function for oxygen O5
corresponds to the electron density map found previously by Superflip. We are making this
check because refinement of sawtooth function for oxygen in the structure containing the
heavy yttrium could be unreliable

Return to the result saved previously (YPO-for-maps):

“File → Structure → Open”: select the name of the previously saved structure (“YPO-for-
maps”)
“File → Structure → Save as”: make a backup copy
Run Refine to re-generate phases
(R factors 11%, 8%, 12% and 20%.)
“File → Structure → Copy in” copy only (!!!) m40 from the final structure
(From this point Refine must not be started)
Run Contour as in the point 12 and draw atomic modulation curves for O5 (using “Atoms
edit” button) in the sections x1-x4, x2-x4 and x3-x4 (difference Fourier)

The following maps will be obtained:

142 Jana2006 Cookbook, version October 2017

By this procedure we plot the final refined modulation curve of O5 together with the
previously calculated Fourier map. The technical trick consists in the fact that Fourier
calculation does not use M40; it uses M80 created by Refine. Only the modulation curve
displayd by Contour is based on parameters fromM40.

143 Jana2006 Cookbook, version October 2017

15. Drawing of refined modulated structure
Return back to the refinement result using “File → Structure → History”

Drawing programs such as Diamond cannot draw a modulated structure directly. Jana2006
creates an approximant based on the refined modulated functions. During this process the
superspace symmetry is used to generate all atomic positions and the symmetry is reduced
to P1. The size of approximant and the starting t value are to chosen to visualize different
portions of the non-periodic matter. In our case the most convenient supercell is 1x8x1 as
follows from the modulation vector being close to (0,3/8,0).

“Plot structure”
In Diamond symmetry operations must not be used for the approximant (no automatic
structure, no supercell, no coordinations …. ). This is because Diamond does not know
modulation displacements of newly created positions.

16. Checking of refinement listing

Start “Edit/View → View of Refine”
Press “GoTo → List of serious warnings”
The listing contains information that modulation of ADP leads to non-positive values for
some values of t coordinate:

Start “Grapht”
Press New/Edit
For “Parameter to draw” select “ADP parameter”
For “Central atom” use “O2” (the label may differ)
For “Principal” select “U(min)”; OK

The following plot shows that U(min) is negative for a small region of t

144 Jana2006 Cookbook, version October 2017

Create the same graphs for O4 and O5

Decreasing of number of ADP modulation waves often solves such situation:

In Atoms/Edit select all oxygen atoms and decrease number of ADP modulation waves to 1
Check refinement listing and Grapht graphs
In our case it only helps for O4. The reason for bad ADP modulation of some oxygen atoms is
probably caused by the fact the sample is more complicated than presented here. High
exposition reveals another set of weak satellites, which need another q-vector for their
indexing. The five-dimensional structure remains unsolved.

145 Jana2006 Cookbook, version October 2017

146 Jana2006 Cookbook, version October 2017
Example 5.2: Na2CO3
Simple modulated structure from single crystal data
Revised 26 September 2017

Na2CO3
Single crystal data measured with KUMA four-circle diffractometer
(KUMA has changed to Oxford Diffraction, later on Varian, later on Agilent, now Rigaku Oxford
Diffraction)
Input files: naco.hkl, naco.cif_od
Frame scaling, absorption correction: done

1. Creating new jobname

Start Jana2006
“File → Structure → New” opens a file manager
Left pane: locate directory with input files
Right pane: double-click naco

2. Import Wizard
Select “Known diffractometer formats”; NEXT
Select “Oxford Diffraction → CCD”; NEXT
[On the screen: “Define basic input file:"]
Select “Input from cif_od file”; file name “naco.cif_od”; OK
[On the screen: “Complete/Correct experimental parameters"]
Leave all settings unchanged; NEXT
q-vector read from the cif_od file should be: 0.1776 0 0.3252

[On the screen: “Define the reference cell/split by twinning"]

Leave all settings unchanged; NEXT

147 Jana2006 Cookbook, version October 2017

 OK to confirm reading of 55555 reflections
For absorption correction select “None or done before importing”; NEXT; FINISH

3. Data Repository
The line in data repository window corresponds to one data set.
OK; YES to accept the data set
[On the screen: “Information about starting symmetry wizard”]
NEXT

4. Symmetry Wizard
[On the screen: “Tolerances for crystal system recognition”.]
Symmetry wizard can be started separately by “File → Reflection file → Make space group
test”.
Change “Maximal deviation for cell angles” to 0.3
Without this change unit cell angles will not be recognized as monoclinic
Leave all other settings default; NEXT
[On the screen: “Select supercell”.]
Leave all settings default; NEXT
[On the screen: Select Laue symmetry]
Select Monoclinic Laue symmetry; NEXT
[On the screen: Select cell centering]
Select C centered unit cell; NEXT
[On the screen: Select space group]
Select C2/m; NEXT
[On the screen: Select superspace group]
C2/m(a0g)0s
[On the screen: Final step of the superspace test]
Accept the superspace group transformed into the original cell: C2/m(a0g)0s; FINISH

5. Creating refinement reflection file

We have finished creation of the file M95 (reflections converted from diffractometer format
to Jana format) and M50 (cell parameters, symmetry). At this step, we will create the file
M90 containing data set merged by symmetry, with forbidden reflections omitted.
NEXT to confirm threshold 3sigma
[On the screen: settings for merging of reflections]
Leave all settings default; NEXT
The Rint for observed reflections will be 4.18%
Accept the data block and calculate coverage; FINISH
[On the screen: Information that structure solution will start]

6. Structure Solution Wizard

[On the screen: window of Structure solution]
In “Formula” textbox type list of chemical elements: Na C O
Select “use Superflip”; “Peaks from Jana2006”

Superflip converges (after noise suppression) with R value 21%. It confirms the symmetry
C2/m(a0g)0s.

148 Jana2006 Cookbook, version October 2017

 This is symmetry found
by Superflip in the
electron density map

This is fit of symmetry

elements supplied by
user (small numbers =
good fit)

Close the listing and press “Accept last solution”

7. Reflection report
Start “Edit/View → View of Reflection report”
Reflection report shows which symmetry operations were used for merging of reflections
and which extinction rules were applied.

8. Verification of the structure solution

Start “Edit/View → editing of M40 file”.
This is the file of structure parameters. Each atom has one modulation wave (number of
waves and six parameters are indicated in red) found by Jana2006 in the electron density
map from Superflip.

Na1 1 1 0.500000 0.670666 0.000000 0.246168 000 0 1 0

0.037995 0.000000 0.000000 0.000000 0.000000 0.000000 0000000000
0.000000 0.066260 0.000000 0.000000-0.004772 0.000000 000000
0.000000

Note that there is no oxygen in M40 – later on some carbon atoms will be changed to
oxygen. The coordinates may differ from the ones shown here dependent on which
equivalent origin position was used by Superflip.
Close the editor
Start “Edit atoms”. It shows the same information in a graphical interface; parameters of
atoms can be accessed by double clicking on atom or by “Action → Edit/Define atoms”;
OK
Start "Plot structure"
Select “Draw average structure”
Start "Draw+continue"
149 Jana2006 Cookbook, version October 2017
 With this option the basic 3d structure will be plotted
In Diamond use Ctrl-A and Delete to remove an automatically created plot (if present)
In Diamond use to include all symmetry independent atoms from M40
Delete Na atoms
In Diamond use to add coordinated atoms
Using mouse cursor and rotations identify atoms which should be oxygens of CO3 group:
C1 and C2
Quit Diamond (without saving the plot)
Start "Edit atoms"
Change chemical type of C1, C2 to oxygen
Use “Action → Rename selected atoms to atom type + number” for oxygens and carbon
The structure now contains: Na1, Na2, Na3, O1, O2, C1
Return to the basic window

9. Refinement of modulated structure with modulations of positions

In Refinement commands set 100 refinement cycles
Run refinement.
Modulation waves are directly refined because the structure found by charge flipping has
already modulation waves. Refinement converges with Robs factors 17%, 13%, 14%, 24%,
30% and 49% for all reflections, main reflections, 1st, 2nd, 3rd and 4th order satellites,
respectively.
Start “Edit atoms” and change ADP of all atoms to harmonic
Repeat refinement
Robs factors: 13%, 8%, 11%, 21%, 26% and 49%
Start “Edit atoms” and add another position modulation wave for all atoms

Repeat refinement
Robs factors: 12%, 7%, 10%, 16%, 23%, 48%
Start “Edit atoms” and add another position modulation wave for all atoms
Repeat refinement
Robs factors: 8%, 8%, 7%, 9%, 15%, 25%
Start “Edit atoms” and add another position modulation wave for all atoms
Repeat refinement
Robs factors: 8%, 6%, 7%, 9%, 14%, 21%
Refinement with four position modulation waves is reasonable because there are satellites
up to the 4th order in data. However, it is not sure at this point that all atoms really need so
many waves.

150 Jana2006 Cookbook, version October 2017

 Use “File → Structure → Save As” and create backup copy “naco_posmod”
(Continue with the original structure)
10. De Wolff sections
In the next step modulation functions will be checked against observed and difference
Fourier map.
Start Fourier commands
For “Map type” select “F(obs) – Fourier”
Switch to the page “Scope”
Select “by a central point”; For “Map axes” select x1 for “1st” column and x4 for “2nd”
For “Center” type Na1
For “Scope [A]” type 2 1 1
Change “Step [A]” from 0.02 to 0.05 to speed-up the calculation

OK; Yes+start
Fourier calculates a four-dimensional parallelepiped x1–x4–x2–x3 with dimensions 2-?-1-1
Å. In x4, the calculation will run over two periods: from 0 to 2 with step 0.02. The step in
other directions will be 0.05 Å. The two-dimensional sections x1-x4 will be stacked along x2;
the three-dimensional "sections" x1-x4-x2 will be stacked along x3.
Start Contour; press “New plot”
Select “General Contour procedure”, “Use old maps” and “Draw maps as calculated”; OK
[On the screen: the first x1-x4 section]
Press “Sum on” and select summation over x2 and x3
This merges all x1-x4 sections stacked along x2 and x3
[On the screen: merged section]
Press “Atom edit”; for “Atom name” type Na1; Add; OK
[On the screen: merged section with refined modulation function of Na1]

151 Jana2006 Cookbook, version October 2017

Please note that in your calculation positions of atoms may be shifted (for instance y of
Na2 from 0.5 to 0).
Press “New plot”
In the next dialogue select “General Contour procedure”, “Calculate new ones” and “Draw
maps as calculated”
[On the screen: Fourier commands, page Scope]
Change “Map axes” to plot sections x2-x4 instead of x1-x4:

“Scope” has been refreshed automatically.

Fourier will calculate a four-dimensional parallelepiped x2–x4–x1–x3 with dimensions
2-?-1-1 Å. In x4, the calculation will run over two periods: from 0 to 2 with step 0.02. The
step in other directions will be 0.05 Å. The two-dimensional sections x2-x4 will be stacked
along x1; the three-dimensional “sections” x2-x4-x1 will be stacked along x3.
OK
[On the screen: the first x2-x4 section]
Press “Sum on” to merge the sections
[On the screen: merged sections x2-x4 with modulation curve of Na1]

152 Jana2006 Cookbook, version October 2017

Press “New plot”
In the next dialogue select “General Contour procedure”, “Calculate new ones” and “Draw
maps as calculated”
[On the screen: Fourier commands, page Scope]
Switch to page “Basic”
For “Map type” select “F(obs)-F(calc) – difference Fourier”
OK
Press “Sum on” and merge the sections
Press “Contours” and define step 0.1 for both positive and negative contours
[On the screen: merged difference section x2-x4]
The difference peaks about 0.5 e-/A3 , indicating modulation of ADP

153 Jana2006 Cookbook, version October 2017

The process could be repeated for other atoms of the structure but usually we can use the
automatic procedure described below. Limitation of this automatic procedure is that all
calculated sections have the same parameters (scope, width, summation …)

Start Contour and press “New plot”

Select “Show de Wolff’s sections”
For “New map to draw” select “F(obs) – Fourier”
Define limits of x4, primary and secondary scope, and interpolation step according to the
following screen shot:

Primary scope means the scope of the coordinate displayed in the plot; secondary scope
means the scope of the other two coordinates. Sections will be automatically summed.
Press “Define atoms for de Wolff’s sections”

154 Jana2006 Cookbook, version October 2017

Using the List button, select all toms, press “Select section for all three coordinates”, and
press “Add”.
OK; OK
Contour calculates the de Wolff section as defined: for each atom sections x1-x4, x2-x4 and
x3-x4 will be calculated, with scope 2 1 1, 1 2 1 and 1 1 2, respectively. The coordinate x4 will
be calculated for two periods with step 0.05.
In Contour, press “W+” “W-“ for displaying the sections.
Repeat the same procedure for difference Fourier sections
Selected sections are displayed here (difference maps use the contour step 0.1):

x2-x4 for Na2: observed and difference Fourier map

155 Jana2006 Cookbook, version October 2017

x2-x4 for Na3: observed and difference Fourier map

x2-x4 for O1: observed and difference Fourier map

156 Jana2006 Cookbook, version October 2017

x2-x4 for O2: observed and difference Fourier map

x2-x4 for C1: observed and difference Fourier map

For C1 no ADP modulation is indicated in the difference Fourier sections

157 Jana2006 Cookbook, version October 2017

11. Refinement of ADP modulations
Start “Edit Atoms”
Select all atoms except carbon and add one ADP modulation wave
The corresponding textbox has label “ADP 2nd” which means ADP described by second order
harmonic tensor

Run refinement
Robs factors: 6%, 5%, 5%, 7%, 12%, 20%
Start “Edit/View → editing of M40 file”.
This is the structure parameter file. Each atom has four modulation waves (number of waves
and 24 parameters indicated in red) and – except carbon – one ADP modulation wave
(number of ADP modulation waves and 12 parameters indicated in blue)

Na1 1 2 0.500000 0.670641 0.000000 0.247961 000 0 4 1

0.019755 0.023192 0.028421 0.000000 0.008258 0.000000 0101111010
0.000000 0.065455 0.000000 0.000000-0.005129 0.000000 010010
-0.001028 0.000000-0.001123 0.001502 0.000000 0.004475 101101
0.000000 0.001068 0.000000 0.000000-0.003381 0.000000 010010
0.001307 0.000000 0.000797 0.000771 0.000000 0.000106 101101
0.000000 0.000000 0.000000 0.000228 0.000000-0.001917 000101
0.000000 0.000000 0.000000 0.004284 0.000000 0.001486 000101
0.000000

Close the editor

Start “Edit Atoms”
Select all atoms except C1 and add another ADP modulation wave
Repeat refinement
Robs factors: 5%, 4%, 4%, 5%, 9%, 14%

According to de Wolff sections, O1 has the second smallest position modulations. Because
ADP modulation usually corresponds to changes in environment of an atom due to position
modulation, we should check whether ADP modulation of O1 is reasonable.

Remove ADP modulation of O1

Run refinement
Add one ADP modulation wave for O1
Run refinement
Add another ADP modulation wave for O1
Run refinement
The first ADP modulation wave of O1 has improved fit significantly. For the second ADP
wave the improvement was not so evident. Therefore, for O1 we shall use maximally two

158 Jana2006 Cookbook, version October 2017

ADP waves.

For the remaining atoms we shall add ADP waves one-by-one to see where to stop.
Add the 3rd ADP modulation wave for Na1
Run Refinement
Robs factors: 4.5%, 3.8%, 3.5%, 5.3%, 8.8%, 14.2%
Add the 4th ADP modulation wave for Na1
Run Refinement
Robs factors: 4.5%, 3.8%, 3.5%, 5.2%, 8.6%, 14.3%
Remove the 4th ADP modulation wave for Na1
Run Refinement

Add the 3rd ADP modulation wave for Na2

Run Refinement
Robs factors: 4.5%, 3.8%, 3.5%, 5.3%, 8.8%, 14.2%
Remove the 3rd ADP modulation wave for Na2
Run Refinement

Add the 3rd ADP modulation wave for Na3

Run Refinement
Robs factors: 4.5%, 3.8%, 3.5%, 5.3%, 8.8%, 14.2%
Remove the 3rd ADP modulation wave for Na3
Run Refinement

Add the 3rd ADP modulation wave for O2

Run Refinement
Robs factors: 4.3%, 3.7%, 3.3%, 5.0%, 8.0%, 13.6%
Add the 4th ADP modulation wave for O2
Run Refinement
Robs factors: 4.2%, 3.8%, 3.2%, 4.6%, 8.0%, 12.0% (for 196 parameters)

Finally, we shall optimize position modulation of carbon:

Remove the 4th position modulation wave for C1
Run Refinement
Robs factors: 4.2%, 3.8%, 3.2%, 4.7%, 8.1%, 12.0%
Remove the 3rd position modulation wave for C1
Run Refinement
Robs factors: 4.9%, 3.8%, 4.0%, 6.0%, 9.9%, 16.0%
Return the 3rd position modulation wave for C1
Run Refinement
Robs factors: 4.2%, 3.8%, 3.2%, 4.7%, 8.1%, 12.0% (for 192 parameters)

159 Jana2006 Cookbook, version October 2017

 This is the final parameters set:

Atom Position waves ADP waves

Na1 4 3

Na2 4 2

Na3 4 2

O1 4 2

O2 4 4

C1 3 0

12. Extinction correction

Start “Edit/View → View of Refine”
Press “GoTo” and select "Statistics Fo, sin(th)/lamba after refinement"

The shell of the strongest reflections has R factor 2.66% while the next strongest shell has
2.32%.
Close the refinement listing
Start “Parameters → Extinction”
Select Isotropic, Type1, Gaussian; OK; Yes to rewrite changes
Run refinement
Robs factors: 4.1%, 3.6%, 3.2%, 4.7%, 8.1%, 12.0% (for 193 parameters)
Check refinement listing

The partial R factor for the strongest reflections is better (2.48%) but still slightly worse than
the neighboring shell.

13. Final check of Fourier sections

Using the same tool like before, plot de Wolff sections x2-x4 through the difference Fourier
map for all atoms of the final structure model. The checkbox “pdate existing sections” must
be selected.
The following sections are displayed for the contour step 0.1. This must be set manually for
each displayed section, otherwise Contour select the contour step automatically.

160 Jana2006 Cookbook, version October 2017

 Na1 Na2 Na3

O1 O2 C1

Use “File → Structure → Save As” and create backup copy “naco_final”

161 Jana2006 Cookbook, version October 2017

14. Plotting of sodium carbonate with Diamond
In this part names of atoms and position of the origin may differ from your results
Start “Plot structure”
In the following dialogue we can either plot the average structure or the approximant
structure. Average structure plot uses positions of the basic structure and 3d symmetry. For
approximant structure, Jana2006 expands the modulated structure to a user-defined area.
Then the expanded structure is exported to the plotting software with P1 symmetry. In the
plotting software (here Diamond) no symmetry operations are allowed including
translations! The reason is that the plotting software cannot transform modulation
displacements.
Select “Draw average structure”; start "Draw+continue"
Use Ctrl-A and Delete to remove an automatically created plot (if present)
Go to “Build → Fill → Super cell” and select “3x3x3” supercell
“Picture → Viewing direction → c”
to plot unit cells
Go to "Build → Add cell edges" to define how many unit cells should be visible
“Objects → Coordinate system” to see axes
By selecting column of atoms and pressing you can create the following figure:

This plot allows for estimation which supercell we shall need for plotting of approximant
structure. We shall use supercell 2 x 3 x 10
Close Diamond without saving
[On the screen: basic window of Jana2006]
Start “Plot structure”
Select “Draw approximant structure” and define supercell 2 x 3 x 10

162 Jana2006 Cookbook, version October 2017

“Keep atom names” means that labels in Diamond will contain the original names from M40
followed by a sequence number”
Start "Draw + continue"

[On the screen: Diamond]

Use to add all atoms of the approximant to the plot. No operation creating new atomic
positions is allowed
“Picture → Viewing direction → c”
to connect atoms
On the right sidebar switch to “Table of atom groups”
Right-click Na, press “Select atoms by group”
In the plot hit the right mouse button on some of the selected sodium atoms
From the context menu Select “Edit → Atom design”
Switch to the page “Radius” and define radius 0.2 Å

163 Jana2006 Cookbook, version October 2017

Using mouse and Delete (on the keyboard) and Ctrl-I (inverts selection) separate the
following motif (three columns of sodium around the column of CO3):

(In case bonds obscure atoms too much, use Ctrl-A + Right click on a selected object + Edit
→ Atom design and change bond style from “thick” to “thin”)
Note starting letters (Na1, Na2 or Na3) of labels of sodium atoms which are not in the
middle of the plot
On the right sidebar switch to the table “Table of created atoms”
Order atoms by name and select the noted sodium atoms
“Build → Polyhedra → Add coordination polyhedra”
Select oxygen for “Ligand atom groups”

OK
With make sure that Rendering and Lightning is activated
The resulting plot will be:

164 Jana2006 Cookbook, version October 2017

Press Ctrl-A to select all atoms
Hit some of the selected objects with the right mouse button
From the context menu select “Edit → Polyhedron design”
Define transparency 0.4; OK

Using rotations and zooming tool create the following figure:

Select some sodium and carbon atoms

Hit one of them with the right mouse button
From the context menu select “Add → Atom labels”
For “Relative position” type “0 0 10”. The last number will move the labels above the
polyhedra
Correct positions of labels with mouse
This is the final plot (labels may differ in your result):

165 Jana2006 Cookbook, version October 2017

 In the middle column the CO3 groups alternate with Na atoms.

15. Properties of CO3 anion

Start commands for Dist (click right mouse button on Dist icon)
In “Basic” page select “d(max) derived from atomic radii and typical distances”
Jana uses typical distances from International Tables. When typical distance is not available,
atomic radius is used instead. The d(max) value is expanded by user defined %. Both typical
distances and atomic radii can be changed in EditM50-Page Composition
For “Listing form”, select “Without symmetry codes”
Press “Define coefficients for bond valences”
[On the screen: dialogue for definition of bond valence constants]
Select “1st atom type” carbon, “2nd atom type” oxygen
Press “From file” and select constants given in the second line
The list displays bond valence constants from various literature sources
Add; OK
[On the screen: Distance commands]
Switch to the page “Select atoms”
Press "Refresh"; for central atom select (with ctrl-left mouse button) C1
For neighbor atoms select O1, O2
Switch to the page “Modulation”
Activate “Make coordination t- tables”
The default parameters for t have the following meaning: distances will be calculated for t
between 0 and 1 with step 0.01, i.e. for 100 values of t. The 100 calculated distances will be
used for calculation of d(min), d(max) and d(ave) printed in the listing. These values will be
followed by list of distances printed for each 10th value of t, i.e. for t between 0 and 1 with
step 0.1

166 Jana2006 Cookbook, version October 2017

OK; YES+Start; YES to open listing
[On the screen: listing of Dist]
Use “Goto → List of distances”
Here the distances are shown as explained previously. The first distance (introduced with
dotted line) is the distance between reference positions of the basic structure, which does
not occur in the real modulated structure
Use “Goto → List of bond valences”
Here minimum, maximum and average BVS for carbon is printed
At the bottom of the listing there is a Coordination table for carbon showing the three C-O
distances for each step in t
Close listing of Dist

Similar information can be obtained from Grapht tool

Start Grapht
Press “New/Edit”
For “Parameter to draw” select “Distances”
For “p” type lower limit 1 and upper limit 1.4
“p” stands for “parameter” to be drawn
For “Central atom” type C1
Press “Define neighbor atoms” and select O1, O2

OK; OK
[On the screen: Plot of three C-O distances as a function o t]
167 Jana2006 Cookbook, version October 2017
As we have already seen in Dist listing, C-O distances are almost constant over the whole
interval of t
Press “New/Edit”
Change “Parameter to draw” to “Bond valence sum”
For “p” change lower limit to 3.5 and upper limit to 4.2
Press “Define bond valence coefficients”
The dialogue shows the coefficients defined in Dist commands are still active
OK; OK
[On the screen: Plot of BVS of carbon as a function of t]
Using the mouse cursor we can find the fluctuation of BVS is between 3.94 and 4.04

Both Grapht and Dist confirm the C-O bonds are constant. Now we shall check planarity of
the CO3 group:
Start "Edit Atoms"
Select atoms O1, O2 and C1
Choose Action "Make symmetrically contiguous motif"
OK; YES to save changes
With this procedure we transformed positions of C1, O1 and O2 to be connected without
need of symmetry transformation
Start commands for Dist
In page “Basic” for “Listing form” check “With symmetry code”
In Page “Select atoms” make sure the central atoms is still carbon and the neighbor atoms
are oxygens
OK; YES+Start; YES to open the listing
“Goto → List of distances”

There are three distances: C1-O1, C1-O2 and C1-O2. For the third distance ,C1-O2, oxygen
O2 must be transformed from the position given in M40 by the symmetry operation x,-y, z,
which is encoded to the Jana symmetry code #s4 (i.e. the fourth symmetry operation). Other
168 Jana2006 Cookbook, version October 2017
distances do not need a transformation because "Symmetry contiguous motif" has been
created in the previous step.
Symmetry codes of may differ for various Superflip runs!
Close the listing of Dist
Start commands for Dist
In page “Basic” press the button “Planes”
For “Atoms defining the best plane” type “O1 O2 O2#s4”
For “Additional atoms” type C1
Add; OK; OK; YES+Start; YES to open listing
“Goto → List of best planes”
The listing shows that the distance of C1 from the plane of oxygens does not change with
the t coordinate which means the planarity of CO3 is not influenced by modulation.

At the next step we will test whether the atoms of CO3 group move during modulation in the
phase, i.e. like a rigid body. This can be visualized by “animated” plot in Contour.
Run refinement (at least zero cycles) to get input M80 for Fourier calculation
Start Fourier commands, select “Map type” “F(obs) – Fourier”
In the page “Scope” select “automatically”, “Use default map orientation”
OK; YES+Start; don’t include new atoms; don’t open listing
Start Contour; New Plot
Select “General Contour procedure”, “Use old map”, “Draw a general section”; OK
[On the screen: form for definition of the general section]
For the first atom type O1
For the second atom type O2
For the third atom type O2#s4
The three atoms define the plane of the section through the electron density map. The
dimensions of the plot are defined in Scope
For “Scope” type “4 4 0”
For “Interpolation step” type “0.02”
This means the plot will be 4x4 Å. Number of parallel sections calculated by the program is
1+s3/is, where “s3” is the width of the section and “is” is the interpolation step. Here it
means one section, because the third parameter is zero.
For “First point put to” type “1 1 0”

The first point of the plot is O1. The horizontal axis of the plot will be O1-O2. The last atom
O2#s4 will be above the horizontal axis.

169 Jana2006 Cookbook, version October 2017

 OK

Click the middle maximum and select all close atoms; OK

Press “Movie”; keep default parameters; OK

The tool plots the successive sections calculated previously by Fourier for selected x4 values.
Sampling in x4 na be changed in Fourier option - page Scope.
ESC to break the movie
Press “t-map ON” and do the same animation for t-map.
For this case geometry of A3-A3 map is similar to that in R3-R3 map. Atoms of CO3 are
moving in phase and probably they could be described by rigid body. The position of one of
oxygens is very stable comparing to the other atoms. This is the oxygen described with less
ADP waves.

16. Sodium atoms

Using Grapht, plot Na-O distances for Na1, Na2 and Na3. In the Grapht dialogue select
“Draw non-modulated curves” – they will appear in the plots like dashed lines indicating
distances in the basic structure.

170 Jana2006 Cookbook, version October 2017

 Distances Na3-O Distances Na2-O

Distances Na1-O

As it has been previously indicated by plot of the structure, Na1 which alternates with
carbon in a column along c, has different coordination comparing with octahedrally
coordinated Na2, Na3. Na1 has more neighbors with strongly varying distances which are all
longer comparing with Na2 and Na3. Good test if this complicated model is reasonable
provides plot of the bond valence sum.

Using Grapht plot graphs of bond valence sums for Na1, Na2 and Na3 (do not forget to
define coefficients for Na-O bond)

171 Jana2006 Cookbook, version October 2017

 BVS for Na3 BVS for Na2

BVS for Na1

Graphs of BVS show that Na1, Na2 are oversaturated while Na3 is slightly undersaturated.
However, the fluctuation of BVS is in all cases small and similar. This means the modulation
of Na3 and its coordination is properly described with approaching and departing oxygen
atoms in balance.

172 Jana2006 Cookbook, version October 2017

Example 5.3.1: CrPO data
Data processing for the room temperature incommensurate phase of Cr2P2O7.
Revised: 30 November 2016

Cr2P2O7
Single crystal data measured with Oxford Diffraction four-cycle diffractometer
Input files: directory CrPO_RT available at http://crysa.fzu.cz/jana/workshops/CrPO_RT.zip

This example is based on the Crysalis 171.38.43g. Crysalis is a program for controll and data
processing of Rigaku Oxford Diffraction (ROD) diffractometers. It has very friendly licencing
policy: everybody can install Crysalis for processing data acquired with a ROD instrument
(which is the case of Cr2P2O7).

1. Getting the file

Start your browser
Connect to http://crysa.fzu.cz/jana/workshops/CrPO_RT.zip
Log in as cookbook with password jana2006
Download the archive CrPO_RT.zip
Unpack the archive
It will create directory CrPO_RT

2. Starting Crysalis
Start Crysalis
In a dialogue that follows locate the file CrPO_228_4.par
Crysalis opens the experiment.
Use to decrease the image size
[On the screen: the basic window of Crysalis]

173 Jana2006 Cookbook, version October 2017

3. Determination of the unit cell

Start Lattice wizard using the button

[On the screen: the window of Lattice Wizard]

In this wizard, each button has a default function started by pressing the centre of the
button and various optional functionalities, which are offered after pressing “>” part of the
button
Start “Peak hunting” with default options
About 6500 peaks have been found
Start “Unit cell finding” with default options
The program finds a unit cell 7.03, 8.42, 4.64, 90, 108.60, 90 and indexes about 30% of
peaks

Start “Ewald Explorer” with default options

Use the buttons to see projections of the found peaks along the unit cell
axes.
Set projection along b*
174 Jana2006 Cookbook, version October 2017
[On the screen: Ewald – projection along b*]

From this projection, it is clear that the additional spots cannot be indexed by any
reasonable supercell.
Set projection along a*
The peaks between the equidistant planes are evidently some spurious peaks arising e.g.
from a small crystal attached on the main sample. We will now remove these peaks
In the right panel, switch to the page “Selection/Lattice-It”
Activate the radio button “Lattice-It”
[On the screen: Ewald Explorer with activated function “Lattice-it”]

The cursor changes to a pair of white lines.

Orient these lines parallel with the horizontal rows of the peaks.
After some delay, the program automatically recognizes the periodicity.
Click the left mouse button to fix the selection. The spurious peaks now become green.
[On the screen: Ewald Explorer with activated function “Lattice-it” and recognized
periodicity]

175 Jana2006 Cookbook, version October 2017

 In the right pane, press “Move selected to:”
This will move the selected (green) peaks to the Group2

At the bottom of the right panel, clear the checkbox of the group 2.
The spurious peaks now become invisible
At the right panel, switch to the page “Crystal”
Select vie along b*
Now we have a clear figure suitable for determination of the q-vector

4. Determination of the q-vector

For selection of the q-vector we should consider the following:
- All satellites must have the same direction with respect to their parent main reflection, and
this direction is defined by the q-vector
- First order satellites are usually stronger than the higher order ones
- The projection of peaks onto a*, b* and c* in the right panel suggests that the q-vector has
the beta component zero
- in order to be consistent with the workshop example 5.3.2, we would like the alpha
component positive (towards the label a*) and with the gamma component negative
(opposite to the label c*). The labels a* (red) and b* (blue) are rather small, located at the
end of the lattice vectors going through the origin.

From these considerations we can suggest a q-vector like this:

176 Jana2006 Cookbook, version October 2017

In the right panel, press “Activate incommensurate peaks” and select “Add 1d modulation
vector (hklm)”
Activate the checkbox close to the q-vector and in the textbox that appears type estimated
components of the q-vector and the satellite index:
0.33 0 -0.5 2

Play with the arrows close to “Modify” button until the satellite position fit with the
projections

177 Jana2006 Cookbook, version October 2017

 Toggle the checkbox close to the q-vector and check if the
peak positions are properly described.
In some areas we can see that there are also the 3rd order satellites:

Using the button “Modify” at the right panel, increase the satellite intex to four.
The reason for four is that when we can see the third order satellites by an eye there might
be the forth order satellites not found by the default peak search

Check views along a* and c* to be sure that the satellites are described
Close the window of the Ewald explorer

5. Refinement of the q-vector

[On the screen: Lattice wizard]
Press the button “Incommensurates/Quasi-crystals”
178 Jana2006 Cookbook, version October 2017
 In the displayed text box, define maximal satellite index (4) and the accuracy (0.05) defining
the maximal tolerancy for an integer (for instance, 1.04 will be rounded to 1)

OK
The program refines the unit cell and q-vector an indexes 93% of the peaks. This is quite
acceptable number because the refinement included also the previously suppressed spurious
peaks.

Press the “>” part of the button “Save information” and select “Save specific peak table
only”. Make backup copy of the peak table under some name which cannot be overwritten
by Crysalis, e.g. “incommensurate”

Close the window of the Lattice Wizard

6. Data reduction
[On the screen: Basic window of Crysalis]
In the right panel, press the button “Start-Stop”
In the next window, select “Data reduction with options”
[On the screen: Step 1: Orientation matrix for data reduction]

179 Jana2006 Cookbook, version October 2017

Check the modulation vector (button “Edit q”) and maximal satellite index (button ”m=”)
The unit cell found previously is C-centered. Now we are having two options: “Use filter for
C-lattice” or “Don’t use filter (P-lattice)”. Using the filter is good for cases with close
diffraction spots because during the data processing, Crysalis finds optimal positions for
diffraction spots and it could skip from a systematically absent satellite position to an
existing position. On the other hand, with less dense diffraction pattern, not using the filter
gives us the complete information and we can verify the C centering. in our example, the
filter for C-centering is not necessary.
NEXT; NEXT
[On the screen: Step 3: Basic algorithm parameters]

Uncheck “Follow significant sample wobbling” (we know that the sample was not wobbling)
Note: In this case, no special parameters are necessary. For more complicated modulated
structures, e.g. with short modulation vector, there are two typical parameters one should
consider: (1) Disabling the option “Refine q-vectors(s)”; (2) “Overriding integration mask
size” available through the button “Edit special parameters”; (2) “Skiping model refinement”
available through the button “Edit special parameters” followed by Alt-E keyboard shortcut.

180 Jana2006 Cookbook, version October 2017

 NEXT;
[On the screen: Step4: Background evaluation]

Keep all settings default; NEXT

The method “Average background” is faster and usually works well; the method “Smart
background” is slower and only sometimes gives better results
[On the screen: Step5: Outlier rejection]
Keep all settings default; NEXT
Outlier rejection based on monoclinic Laue symmetry does not mean that the program
would create a false monoclinic data for non-monoclinic case. Although the algorithm
behind is undescribed, the process is quite conservative and rejects only few reflections. The
resulting data are unmerged.
[On the screen: Step6: Output]
Keep all settings default; FINISH
Following is the data processing. In the first step, the program goes to the peaks positions
predicted by the orientation matrix and make profile analysis to get accurate peak positions
and some other hidden information. During profile analysis is accompanied by the
background subtraction. In the second step the program makes data integration.

7. Output
1) The resulting data are placed to the directory struct\tmp. For Jana2006 we need the files
CrPO_228_4.cif_od and CrPO_228_4.hkl

2) In the left toolbar of Crysalis, there is a button Jana. The button triggers the following
actions: directory struct\ jana_CrPO_228_4 is created; files CrPO_228_4.cif_od and
CrPO_228_4.hkl are copied to this directory; Jana2006 is launched

3) In the left toolbar of Crysalis, press the button

[On the screen: Inspect data reduction results]
Go to the page “Data reduction output”

181 Jana2006 Cookbook, version October 2017

This page shows the merging R values (Rint) sorted by intensity shell. When scrolled up, it
also shows the same information separately for satellites:

The Rint is calculated together for observed and unobserved reflections. In order to see Rint
separately for observed and unobserved reflections, Jana2006 must be used

Go to the page “Graphs”

The most important is the graph from the frame scaling. Too large fluctuation of this graph
usually mean that an absorption correction by shape should be done.

182 Jana2006 Cookbook, version October 2017

 4) Note about the absorption correction
The data were automatically corrected using the empirical absorption correction, called
ASSPACK in Crysalis. In order to improve the absorption correction the following steps should
be done:
- At the bottom of the “Inspect data reduction results”, press the button “Abs display”. This
opens a tool for indexing of the crystal shape
- At the bottom of the “Inspect data reduction results”, press the button “Refinalize” and
repeat the corrections.
Both tools offer rich possibilities and cannot be described within this basic example.

8. Application of the created data set

The created data set could be used for the next example 5.3.2 - solution of the
incommensurate phase of Cr2P2O7. However, in such case Superflip may return different
origin positions than expected in the Cookbook, because the fixed seed of Superflip will not
fix the result when the input data are changed. For this reason, example 5.3.2. is based on
the pre-prepared data set, and the data set from this example is tested at the end.

183 Jana2006 Cookbook, version October 2017

184 Jana2006 Cookbook, version October 2017
Example 5.3.2: CrPO
Solution of the room temperature incommensurate phase of Cr2P2O7.
Revised: 19 April 2017

Cr2P2O7
Single crystal data measured with Oxford Diffraction four-cycle diffractometer
Input files: crpo.hkl, crop.cif_od
Frame scaling, absorption correction: done with software of the diffractometer

1. Creating new jobname

Start Jana2006
“File → Structure → New” opens a file manager
Left pane: locate directory with input files
Right pane: double-click crpo

2. Import Wizard
Select “Known diffractometer formats”; NEXT
Select “Oxford Diffraction – CCD”; NEXT
In “Define basic input file“ select “Input from cif_od file“; OK
Leave all settings unchanged; NEXT
Leave all settings unchanged; NEXT
For absorption correction select “None or done before importing”; NEXT
FINISH
[On the screen: Data repository]
OK; YES to accept the data set

3. Symmetry Wizard
Use default tolerances for crystal system recognition
Select monoclinic Laue symmetry
Select C centering
Select space group C2/m
Select superspace group C2/m(a0g)0s
Accept the superspace group in the original cell

4. Creating refinement reflection file

Follow default steps of the wizard

5. Structure solution wizard

Type the formula Cr2 P2 O7, two formula units
Select “Superflip” and “Peaks from Jana2006”
“Run solution”
Superflip converges with R around 19% and confirms the selected superspace group
Close the listing of Superflip
Press “Accept last solution”

6. Investigation of charge flipping results

185 Jana2006 Cookbook, version October 2017

Start EditM50, page Composition; press “Formula from M40”
The formula based on structure model is Cr2P2O8, i.e. one oxygen more
Start Options for Dist; for “Listing form” select “With symmetry codes”; Enlarge the d(max)
as derived from atomic radii and typical distances to 15%;
Run Dist and open listing
In the listing make sure that the atoms P1, O1, O2 and O3 make symmetrically contiguous
motif, i.e. that O1, O2 and O3 listed in coordination of P1 have one of their symmetry codes
equal to “x,y,z”.
In this example, oxygens are NOT in correct symmetry positions and transformation is
necessary.
Start "Edit atoms"
Select P1, O1, O2, O3
"Action → Make symmetrically contiguous motifs"
Repeat calculation of distances with symmetry codes.
The listing of Dist should look as follows:

The symmetry codes will be different for different Superflip solutions, note your codes!
Start Contour and New plot
Select “General Contour procedure”
Select “Calculate new ones”; OK (if the option is not active, Contour has no old map and it
will calculate a new map by default)
In Fourier option, page Basic, select “Map type” = “F(obs) – Fourier”
In page Scope set calculation of the x1-x4 section with the central point Cr1

OK;
[On the screen: Contour window with plot of x1-x4 section through position of Cr1]
Press “Atoms edit” and add Cr1 to the list of atoms to be plotted
For making de Wolff sections of other atoms, we will use the automatic tool:

186 Jana2006 Cookbook, version October 2017

Start Contour, “New plot”
Select “Show de Wolff‘s sections”
Define parameters of the sections and atoms to be plotted

OK to run the tool

Use “W+” and “W-“ to see the resulting plots

187 Jana2006 Cookbook, version October 2017

 Cr1: section x1-x4 P1: section x2-x4 O1: section x2-x4

O1: section x1-x4 O2: section x1-x4 O3: section x2-x4, scope along x2 is 4 Å

From the plots we can conclude:

(Positions of the centers depend on the Superflip solution and may be different, note your
values! For instance, Cr might have the centre of its modulation function at 0.25)
Cr1 needs a saw-tooth function with center at 0.75 and width 1
P1 needs more modulation waves
O1 seems to be properly described in x2-x4 section but not in x1-x4
O2 needs saw-tooth function with center 0.75 and width 1
O3 needs a crenel function with centre 0.75 and width 0.5;. the remaining maxima could be
described with modulation function related by symmetry operation -x, -y+1, -z+3, i.e.
#s3t0,1,3 in Jana encoding (This can be found from the Dist listing. Again – this operation
depends on the Superflip solution!)

188 Jana2006 Cookbook, version October 2017

 The saw-tooth shape of modulation function van be modeled by three ways:
1) As a discontinuous positional modulation named in Jana2006 saw-tooth positional
modulation
2) As a discontinuous occupational function called crenel function combined with Legendre
position modulation
3) As a discontinuous occupational function called crenel function combined with x-harm
positional modulation.
The first method is obsolete. Here we will use the second method, i.e. crenel function (a
straight line along x4) rotated to the saw-tooth direction by Legendre position modulation.

7. Setting crenel functions

Use “File → Structure → Save As” and create backup copy “crpo_superflip”
Do not continue with the new structure
Start “Edit atoms”
Double-click Cr1
[On the screen: Atom Edit tool]
In page “Define” activate “crenel” and define delta=1 and x40=0.75;
Leave refinement boxes clear (we don’t want refinement of x40 or delta, as x40 is fixed to
0.75 by symmetry and the atom seems to be present for all x4 values); OK
Refinement keys of delta and x40 are not automatic
Select “Legendre polynomials in crenel interval”;
OK
Repeat the process for O2 (here delta=1, x40=0.75)
Repeat the process for O3 (here delta=0.5 and x40=0.75)
Leave “Edit atoms”, YES to save changes
After activation of Legendre polynomials, the keys “Legendre” will appear in M40 file.

Start EditM50, page Composition; press “Formula from M40”

The program now prints correct composition for Z=2: Cr2P2O7. Number of oxygens has been
reduced using crenel function with delta=0.5 for O3
Start Contour, “New plot”
Select “Show de Wolff‘s sections”
Plot again the de Wolff sections for all atoms
The electron density will be still the one from Superflip because we did not refine the
structure. The displayed modulation function will have their initial shape (before
refinement). In case some crenel modulation function does not fit with the corresponding
maximum, check its centre. The most probable reason is that the actual position of the
centre differs of the value given in this cookbook, because Superflip put the origin to a
different equivalent point.

189 Jana2006 Cookbook, version October 2017

 Cr1 O2 O3

For O3 we also display the modulation function related by symmetry operation O3#s3t0,1,3
(or your actual symmetry operation). To display such plot, use the following steps:
Press the Contour button “From list”
Select the section O3_y; OK
Press Atom edit and add O3#s3t0,1,3; OK

Run refinement
Refinement converges with Robs factors 21%, 10%, 19%, 54%, 65% and 129%
Run Contour and plot again the three sections:

Cr1 O2 O3

190 Jana2006 Cookbook, version October 2017

 The coefficients of Legendre polynomials are calculated such that the resulting modulation
function fits best with the original modulation function. However, the first modulation wave
expressed by Legendre polynomials has only the linear (coefficients 1-3) and quadratic
(coefficients 4-6) part so the shape cannot fit exactly.
Start Edit atoms and change ADP of all atoms to “harmonic”
Repeat refinement
Refinement converges with Robs factors 19%, 7%, 20%, 53%, 63% and 123%

Start Edit atoms and add another one position modulation wave for all atoms
Repeat refinement
Refinement converges with Robs factors 11%, 4%, 13%, 20%, 57% and 97% for 68
parameters
Make backup copy “crpo_initial_crenel”

8. Setting up a constraint for discontinuity points of O2 and O3

The shape of the PO4 tetrahedron cannot be changed too much by modulation because it
would contradict the chemistry. Thus the discontinuity points of crenel functions of oxygen
atoms must have the same t coordinate.
Start “Grapht”
Press “New/Edit”
For “Parameter to draw” select “Distances”
For “Central atom” type P1
Step in t 0.001; For “p” type “1.2” and “1.8”;
“p” determines the lower and upper limits of the visualized parameter
Press “Define neighbor atoms” and select only O2 and O3
OK

The resulting plot shows distances P1-O2, P1-O2#s4t0,1,0, P1-O3 and P1-O3#s3t0,1,3 as a
function of t and it is obvious that discontinuities occur for slightly different values of t. This
191 Jana2006 Cookbook, version October 2017
 is however not possible because PO4 cannot be so much distorted. The discontinuous
changes of oxygen positions must therefore occur at the same t value.

The centre x40 of the crenel function of O3 must be fixed at 0.75 (or equivalent value
following from your structure solution), because this atom is in a special position; for
different center, there would be an overlap with symmetry related O3#s3t0,1,3 leading to a
disordered structure. The t-coordinate of this centre is t40[O3].
For O3 delta=0.5, therefore one of the two discontinuity points is t40[O3]+0.25
For O2 delta=1.0 and one of the two discontinuity points is t40[O2]-0.50
The following equation will align the two discontinuity points:
t40[O3]+0.25 = t40[O2]-0.50 => t40[O2]= t40[O3]+0.75 => t40[O2]= t40[O3]-0.25
(The symbol t40 means x40 transformed to t coordinate.) In the last step the equivalent
translation has been applied to get t40[O2] from the interval (0,1).

Start Refinement options, page “Various”

Press “Equations” and add equation t40[O2]=t40[O3]-0.25
Add; OK; OK; YES+Start
Refinement converges with Robs factors 11%, 5%, 12%, 19%, 56% and 104%; i.e. with very
small changes
Start “Grapht” and repeat the same plot like before

The discontinuity points for O2 and O3 now occur at exactly the same value of t
Make backup copy “crpo_aligned”

9. Crenel function for phosphorus

Start Contour and plot the observed Fourier section x2-x4 for P1 using the scope 0 2 0. Plot
similar section for the difference Fourier map.

192 Jana2006 Cookbook, version October 2017

Although P1 seems to be correctly described with harmonic functions, a careful analysis
reveals fluctuation of densit along the modulation function. For this reason, we shall
describe P1 similarly like in case of O3: with two symmetry equivalent crenel functions of the
width 0.5
Quit Contour
Start “Edit atoms”
For P1 activate crenel; define x40=0.86 (as deduced from the Fourier map) and delta=0.5;
activate Legendre polynomials;
Return to Edit page
change y from 0.5 to 0.48 (this might be already done automatically by activation of
Legendre); change site occupancy from 0.5 to 1 (reduced occupancy is now managed by the
crenel function)
Close “Edit atoms”; check formula in EditM50 (should be Cr2 P2 O7 for Z=2)
Start Contour; use old maps
In “Atoms edit” of Contour add P1 and the symmetry equivalent P1#s4t0,1,0

193 Jana2006 Cookbook, version October 2017

 or
The left image is valid for the manual change of y[P1] to 0.48 while the right one arises
when y[P1] is adjusted automatically during the conversion of the harmonic modulation
function to Legendre polynomials.

Quit Contour
Start Refinement option, page “Various”, button “Equations”
Define equation t40[P1] = t40[O3]
This bounds discontinuity points of P1 with O3
Run refinement
Refinement converges with Robs factors 10%, 5%, 10%, 14%, 56% and 96% (77 parameters),
i.e. better fit for 1st and 2nd order satellites
Make backup copy (rewrite crpo_aligned)

10. More modulation waves

At this point all atoms have two position modulation waves and no modulation of ADP. We
will add position and ADP modulation waves until they improve R factors. Normally this
would be result of trials and errors, here the process will be simplified.

Oxygen O1 is only slightly modulated and does not need more waves.

Add another one position modulation wave for Cr1, P1, O2 and O3
Refine
Refinement converges with Robs factors 9%, 5%, 10%, 7%, 58% and 85%
Add the first ADP modulation wave for Cr1, P1, O2 and O3
Start Refinement options, change number of refinement cycles to 100 and set damping
factor 0.4
Refine
Refinement converges with Robs factors 7%, 4%, 6%, 12%, 38% and 57%
Add another ADP modulation wave for Cr1, P1, O2 and O3
Refine
194 Jana2006 Cookbook, version October 2017
Refinement converges with Robs factors 5%, 3%, 5%, 10%, 24% and 39%
Add another position modulation wave only for Cr1
Refine
Refinement converges with Robs factors 5%, 3%, 5%, 7%, 18% and 38%
Add another ADP modulation wave only for Cr1
Refine
Refinement converges with Robs factors 4%, 3%, 4%, 6%, 18% and 34% with 176 refined
parameters.
Make backup copy: crpo_simple

The result obtained here is not bad but a question remains if the poor fit of the 3rd and 4th
order satellites is caused by structure model or weak satellite intensity. For 3rd order
satellites there is still 50 % of reflections above the observability limit. GOF 3.42 also
suggests the model could be improved.

Start “Edit/View → View of Reflection report”

Press “GoTo” and move to “Summary after averaging”
Here we can see Rint for the 3rd order observed satellites (11%) and for the 4th order
observed satellites (22%). This suggests the quality of high order satellites is not so bad.

Start Contour and plot the following difference Fourier sections: x1-x4 for Cr1, x2-x4 for P1,
x1-x4 for O2 and x2-x4 for O3. For plotting use always the contour step 0.2.

195 Jana2006 Cookbook, version October 2017

 Cr1: x1-x4 P1: x2-x4 O2: x1-x4

O3: x1-x4 O3: x2-x4 O3: x3-x4

11. Disordered structure

The difference Fourier sections for Cr1, O2 and O3 indicate a disorder close to the
discontinuity points of crenel function. At this part we shall create a structure model with

196 Jana2006 Cookbook, version October 2017

initial modulation functions describing the disorder.

For each of the atoms delta will be shortened and additional atoms with small
complementary delta will be placed to the difference maxima found in the Contour sections.
The resulting structure model must keep formula unchanged. The delta of all disordered
atoms will be kept the same.

The atoms Cr1 and O3 have to be duplicated and the atoms O2 triplicated. Then the atomic
parameters of the original and split atoms must be modified to account for the disordered
model as found from difference Fourier maps.
Start “Edit atoms”
Select Cr1, O2 and O3; Press “Action” → Select “Split atom”;
Choose 0 for relative occupancy of split atom;
Check the manually defined shift of atoms equal 0 0 0;
Uncheck “Reset ADP to isotropic”;
Uncheck “Generate restrict command for refine;
Press “Apply->Go to next”;
Repeat this procedure for O2 and O3;
Select the atoms O2’ and repeat the split procedure once more to create O2’’;
OK; Yes to the question about rewriting;

The new positions will describe disorder at the discontinuity area. For Cr1’ and O3’ two
positions are generated by symmetry; for O2 we have introduced both positions.

Start “Edit atoms”

Select Cr1’, O2’, O2’’, O3’
Change number of their position and ADP modulation waves to zero
Change ADP to isotropic
For Cr1’ change x to 0.06 and site occupancy to 0.5
For O2’ change x to 0.12 and site occupancy to 0.5
For O2’’ change x to 0.26 and site occupancy to 0.5
For O3’ change z to 1.44 and site occupancy to 0.5
New positions can be found on the above shown difference sections.

Start Refinement option, page “Various”

Press “Equations” and define the following equations:
t40[Cr1’]=t40[Cr1]+0.5
This puts the additional atom Cr1’ to proper position along x4 related to Cr1
delta[Cr1’]=1-delta[Cr1]
This defines delta of the additional atom like complementary to the original atom
t40[O2’]=t40[O3]+0.25
This puts the additional atom O2’ to proper position along x4 related to O3.
t40[O2’] = t40[O2]-0.5 but we can substitute for t40[O2] from already defined equation
t40[O2] = t40[O3]+0.75
delta[O2’]=delta[Cr1’]
We want delta the same for all additional atoms
t40[O2’’]=t40[O3]+0.25
delta[O2’’]=delta[Cr1’]
197 Jana2006 Cookbook, version October 2017
delta[O2]=1-delta[Cr1’]
t40[O3’]=t40[O3]+0.25
delta[O3’]=delta[Cr1’]
delta[O3]=0.5-delta[Cr1’]
After applying these equations all delta intervals will depend on delta of Cr1 and the total
occupation of atoms will be the same like in the simple model.
In page Basic set zero refinement cycles
Run refinement
The refinement applies the equations and calculates structure factors
Because delta[Cr1] is one all delta parameters of additional atoms are zero and R values are
as before: 4%, 3%, 4%, 6%, 18% and 34%
Start EditM50, page “Composition”, and press “Formula from M40”
This shows correctly Cr2 P2 O7

We want to assign some non-zero delta for additional atoms and view modulation functions
but with the original electron density
Start “Edit atoms” and change delta of Cr1 from 1 to 0.9
In refinement we have still zero cycles
Run refinement
This recalculates delta based on equations and creates new (unwanted) input for Fourier
Start “File → Structure → Save As” and create copy “tmp”
Do not continue with the new structure
Change delta of Cr1 back to 1
Run refinement (with zero cycles)
This creates good input for Fourier
Start “File → Structure → Copy in” and copy to the current structure only M40 from the
previously created backup copy “tmp”
Now we have old input for Fourier and new input for plotting modulation functions
Start Contour
For Cr1 plot x1-x4 difference section and visualize Cr1, Cr1’, Cr1’#s2t0,0,2
For O2 plot x1-x4 difference section and visualize O2, O2’ and O2’’
For O3 plot x3-x4 difference section and visualize O3, O3’ and O3’#s2t0,1,3

198 Jana2006 Cookbook, version October 2017

 Cr1: x1-x4 O2: x1-x4 O3: x3-x4

Make backup copy: crpo_disorder_initial

12. Refinement of disordered model

Refinement of the disordered model is susceptible and must be started from very small delta
of Cr1’. The purpose of this refinement is not to obtain proper geometry of the disordered
atoms because the corresponding scattering power is very small: for O3 and delta=0.1 the
disordered atom would have 0.1*0.5*8 electrons, which is less than 0.5 of hydrogen.

Start “Edit Atoms”

Change delta of Cr1 to 0.98 and leave it fixed
Run zero refinement cycles
This will create delta=0.02 for additional atoms
Test Formula (EditM50 – must be Cr2 P2 O7)
Start Refinement options, set 100 refinement cycles with damping 0.4
Run refinement until convergence
The simple model without disorder converged with Robs factors 4%, 3%, 4%, 6%, 18% and
34% for 176 refined parameters.
Now we have obtained Robs factors 3%, 2%, 3%, 4%, 10% and 27% for 192 refined
parameters. This is convincing improvement.
Start “Edit atoms” and activate refinement of delta for Cr1
Run refinement
Refinement converges with Robs factors 2.7%, 2.3%, 2.2%, 3.7%, 9.2% and 18.3% for 193
refined parameters. Delta of additional atoms refines to 0.093.
Start “Parameters → Extinction” and activate isotropic extinction correction
Run refinement
Refinement converges with Robs factors 2.6%, 2.1%, 2.2%, 3.8%, 9.2% and 18.3% for 194
199 Jana2006 Cookbook, version October 2017
 refined parameters.
Start “Edit atoms” and add one position modulation wave to all additional atoms (Cr1’, O2’,
O2’’ and O3’).
Run refinement
Refinement converges with Robs factors 2.4%, 1.9%, 2.0%, 3.5%, 7.3% and 12.2% for 218
refined parameters. GOF = 1.97. Delta of additional atoms refines to 0.105.

Make backup copy: crpo_disorder_final

13. Conclusions
The disordered model does not provide proper geometry of disordered atoms. However, we
can draw an approximant structure in large area and try to determine in which places of the
real space the structure becomes disordered.

Plot the structure expanded to 3x3x3 cells

Indicate the atoms with ‘ and ‘‘ in yellow

In view along c we can distinguish columns without disorder and columns with disorder

Open the structure “crpo_simple”

Start Grapht, press “New/Edit”
Change step of t to 0.0001
Select “Parameters to draw” – “Distances”
Change limits of “p” to 1.2 (lower bonding limit) and 1.8 (upper bonding limit)
Select P1 for central atom
Select O1, O2, O3 for neighbors
OK
200 Jana2006 Cookbook, version October 2017
The definition interval is limited to 0.5 because P1 has a crenel.
The distances at the end of the definition interval are unreasonable
We can read from the plot that the definition interval is approximately tmin=0.21, tmax=0.7

Press “New/Edit”
Change Minimum and Maximum for t: from 0.21 to 0.7
OK

From refinement of disordered structure we have found delta of the disordered area 0.105.
In order to get reliable distances from the simple model we limit the interval where distances
are plotted to reliable interval.

Press New/Edit
Change Minimum and Maximum for t: from 0.21+0.05=0.26 to 0.7-0.05=0.65
OK

201 Jana2006 Cookbook, version October 2017

 In the interval found from refinement of disordered structure the distances are correct.

14. Testing of data prepared in Example 5.3.1.

This example was based on data set prepared independently on the data obtained in
example 5.3.1 in order to get the expected origin position from Superflip. Here we will test
that data obtained from the Example 5.3.1 can be used for the structure model of Cr2P2O7.
Open crpo_disorder_final in Jana2006
Start “File → Reflection file → Import/modify the reflection file”
Start “Reimport”
[On the screen: Data reduction file from:]
Start Browse and localize the hkl file created in Example 5.3.1
This file should be in directory struct\tmp of the Crysalis session
Continue steps of the wizard, accept changes of data repository, follow steps to prepare
refinement reflection file
Run refinement
After initial large change of the scale factor, resulting R values should very similar to those
obtained in this example

202 Jana2006 Cookbook, version October 2017

Example 5.4: Fresnoite
Solution of (3+2)-dimensional incommensurately modulated structure.

Ba2TiSi2O8

Single crystal data measured with Oxford Diffraction four-cycle diffractometer

Input files: fresnoite.hkl, fresnoite.cif_od
Frame scaling, absorption correction: done with software of the diffractometer

1. Creating new jobname

Start Jana2006
“File → Structure → New” opens a file manager
Left pane: locate directory with input files
Right pane: double-click fresnoite

2. Import Wizard
Select “Known diffractometer formats”; NEXT
Select “Oxford Diffraction – CCD”; NEXT
In “Define basic input file“ select “Input from cif_od file“; OK
Change all q-vector components to 0.302 keeping the original sign; NEXT
Leave all settings unchanged; NEXT
For absorption correction select “None or done before importing”; NEXT
FINISH
[On the screen: Data repository]
OK; YES to accept the data set

3. Symmetry Wizard
Use default tolerances for crystal system recognition
Select tetragonal Laue symmetry 4/mmm
Press Show/modify X centering
[On the screen: select individual centering vectors]
Select centering vectors: (0 0 0 0 0) and (0 0 1/2 1/2 1/2); OK; Next
Select space group X4/mbm and note the non-centrosymmetric possibilities
Accept the superspace group in the original cell
For five-dimensional structures the wizard does not offer selection of superspace group. It
will be done later manually.

4. Creating refinement reflection file

Follow default steps of the reflection wizard
The program discards 54372 reflections, 455 observed, Rint=3.28%

5. Structure solution wizard

ESC

6. Solution of average structure – centrosymmetric solution

Superflip for five-dimensional structure returns five-dimensional electron density map.

203 Jana2006 Cookbook, version October 2017

However, Jana2006 would not evaluate initial modulation functions from this map and we
would get average structure. For this reason solution will be done separately for average
structure
Start “Tools → Transformation → Go to average 3d structure”
For name use “ave_centro”
Follow steps of the Reflection wizard
Continue with the new structure
Start “Run → Solution”
For Formula type “Ba Ti Si O”
Select “Superflip” and “Peaks from Jana2006”
Select “Use a specific random seed” and type “111” for random seed
With this option result of charge flipping will be the same in most computers and origin
position and special atomic position will be equivalent with this cookbook. For normal work
Random seed should not be fixed.
OK
Superflip converges with R around 18% and confirms the selected space group
Accept the result
Start EditM50 and check formula for Z=4: Ba2 Ti Si2 O4 (should be Ba2 Ti Si2 O8)
Run Refinement
Refinement converges with Robs=14%
Using “Edit Atoms”: check Uiso of O1 is it rather high ~ 0.13 Å2; Esc
Plot the average structure; Check that the O1 atom is in a close contact with Ba1 ~ 1.16 Å
Run “Edit atom” and delete the atom O1
Run refinement
Refinement converges with Robs=13%
Run difference Fourier
Add difference maxima Max1, Max2, Max3 as oxygens
Run refinement
Refinement converges with Robs=12%
In EditM50 check formula: Ba2 Ti Si2 O8
Run difference Fourier; Add difference maxima Max1 as a new oxygen atom
Run refinement
Check formula: Ba2 Ti Si2 O16
In EditM40 allow refinement of atomic occupancies – “ai” parameter – of all oxygen atoms;
Refinement converges with Robs=9.1%
Check formula: Ba2 Ti Si2 O9.2
Using “Edit atoms” change all ADP to harmonic
Run refinement
Refinement converges with Robs=5.9%
Check formula: Ba2 Ti Si2 O10.1
Plot average structure; make Si-O coordination; draw the projection along [0,0,1]and along
[1,-1,0]

Jana2006 sends to Diamond expanded structure in P1 symmetry. This is because of

centering X which Diamond cannot use. In the plot the silicon atom is surrounded by eight
partially occupied oxygen atoms. This can be either reality which would be described by

204 Jana2006 Cookbook, version October 2017

 crenel functions or wrong symmetry for average structure.

The figures indicate a disorder of SiO4 groups by the mirror perpendicular to the c axis. The
disorder can be described alternatively by the two fold axis perpendicular to c and a+c.
The disorder in the average structure can be produced either by the fact the structure is not
centrosymmetric or by modulation. First we shall test non-centrosymmetric description.

7. Solution of average structure – non-centrosymmetric solution

In order to see all highest non-isomorphic subgroups we shall create a testing structure:
Start “File → Structure → save as” and create a testing structure named “ave-test”.
EditM50: change symmetry from X4/mbm to P4/mbm; Accept changes without re-creation
of the diffraction file;
Start “Tools → Transformations → Go to subgroup structure”; Press the button “Select non-
isomorphic subgroup:”

The highest subgroups not giving a split model for SiO4 group are P-421m and P4bm.
Now we have to check if basic positions (x,y,z) of atoms Si1, O1, O2 and O4 belong to a
unique tetrahedron.
Return to the structure “ave-centro”; Set distance commands: in the “Basic” define
calculation of angles and the listing form “with symmetry codes”.

205 Jana2006 Cookbook, version October 2017

Run dist; Open the listing:

The blue lines indicate the bond distances and red lines bond angles of the basic SiO4 group.
From the symmetric codes we can see that only the atom O2 should be transformed by the
symmetry operation #s10 i.e. by x,y,-z
Start “Edit atoms”: select the atom O2 and transform it by x,y,-z
Before testing of non-centrosymmetric space groups we have to reset the occupancies to
their the values corresponding the SiO4 split. This means that their full values are to be
reduced by 1/2. The atom O3 is coordinating Ti and it can also be shown that it has to
represent a split position.
Start “Edit atoms” set “ai” for O1, O2, O3 and O4 to 0.25, 0.5, 0.25 and 1, respectively.

In the next step we shall test non-centrosymmetric space group X-421m

Start “Tools → Transformation → Go to subgroup structure”
[On the screen: Define subgroup]
The button for selecting non-isomorphic subgroup is missing because of non-standard

206 Jana2006 Cookbook, version October 2017

 centering
Select symmetry operator for the screw axis 21 and rotation axis -4 of the space group
X-421m
Press “Complete subgroup”
The subgroup should be X-421m without changing axes
Next; Next
[On the screen: select one representative of the 2nd coset]
Select inversion as operator for expansion to the subgroup; Next
Deselect “Make as twinned structure”
Type name of the structure (for instance “ave_-421m”); Next
Follow steps of the Reflection wizard
Continue with the new structure
Check formula – still Ba2 Ti Si2 O16 for Z=4
“Edit atoms”: delete atoms O1_2, O2_2 and O4_2 and reduce the occupancy of O3 to 0.25.
Check formula – still Ba2 Ti Si2 O8 for Z=4
In “Refine commands” set damping factor to 0.5
Run Refinement
Refinement converges with Robs=8.3%

In the next step we shall test non-centrosymetric space group X4bm

Start “File → Structure → Open” and return to the centrosymmetric average structure
(“ave_centro”)
Start “Tools → Transformation → Go to subgroup structure”
[On the screen: Define subgroup]
The button for selecting subgroup is missing because of non-standard centering
Select symmetry operator for glide plane b and rotation axis 4 of the space group X4bm
Press “Complete subgroup”
The subgroup should be X4bm without changing axes
Next; Next
[On the screen: select one representative of the 2nd coset]
Select inversion like operator for expansion to the subgroup; Next
Deselect “Make as twinned structure”
Type name of the structure (for instance “ave_4bm”); Next
Follow steps of the Reflection wizard
Continue with the new structure
Check formula – still Ba2 Ti Si2 O16 for Z=4
“Edit atoms”: delete atoms O1_2, O2_2, O3_2 and O4_2.
Check formula – still Ba2 Ti Si2 O8 for Z=4
Run Refinement
Refinement converges with Robs=2.27%
“Edit atoms”: rename oxygens to O1, O2, O3, O4 using “Rename selected atoms to atom
“File → Structure → Save As”: ave_final

8. Back to the modulated structure

“File → Structure → Open”: open “fresnoite”
EditM50:
In “Composition page” type formula Ba Ti Si O and formula units 4
207 Jana2006 Cookbook, version October 2017
 In “Symmetry” page change symmetry to X4bm
OK to confirm centering
OK for the first translation part
OK for the second translation part
Close Editm50
Re-create refinement reflection file
“File → Structure → Copy in”: import only M40 from the average structure "ave_final"
Run refinement
Refinement confirms R value for the average structure:

9. Superspace symmetry
In the previous paragraph the program asked for completion of the translation part of
superspace symmetry operators. For the present case two symmetry operators should be
completed:

and

Each translation part has two elements which can be either 0 or 1/2. Thus we have sixteen
combinations: 0 0 + 0 0, 0 0 + 1/2 0, 0 0 + 0 1/2, 0 0 + 1/2 1/2 … However, the translation
part can be combined with the X centering (0 0 1/2 1/2 1/2) which makes equivalent for
instance the pairs 0 0 + 1/2 1/2 and 1/2 1/2 + 0 0. Moreover, some combinations are
forbidden because they lead to inconsistent translation (0 0 0 1/2 1/2).
Finally we have only two independent possibilities:
0 0 + 1/2 1/2
00+00
208 Jana2006 Cookbook, version October 2017
 Start EditM50
Go to page “Symmetry”
Focus the textbox “Space group” and press TAB. The program offers the first symmetry
operator needing to complete.
At the first attempt try combination 0 0 for the first operator and 1/2 0 for the second one.
The program will not allow it.
Then try the combination 0 0 + 0 0. Close EditM50 and re-create refinement reflection file.
From the screen output of the symmetry wizard note number of rejected reflections: 54372
rejected reflections from which 455 is observed with the largest I/sigma(I) around 15
Return to EditM50
TAB in the textbox “Space group”
For translation parts use combination 0 0 + 1/2 1/2
Re-create refinement reflection file
From the screen output of the symmetry wizard note number of rejected reflections: 55353
rejected reflections from which 463 is observed with the largest I/sigma(I) around 15
This means: the translation part 1/2 1/2 of the second operator leads to 1000 more forbiden
reflections but the number and sigma level of the observed forbidden reflections remains
almost the same like with the translation part 0 0. This confirms the presence of the
translation part 1/2 1/2 for the second symmetry operator.

10. Refinement of modulated structure – position modulation

Make backup copy: “fresnoite_backup”
Start “Parameters → Options”
Press “Define wave vectors”
For the third wave define 1 1
This means modulation wave vector given by a sum of the q vectors
For the fourth wave define -1 1
This means modulation wave vector given by a difference of the q vectors
“Edit atoms”: set two position modulation waves for all atoms
This means one harmonic wave along the first wave vector (q1) and one harmonic wave
along the second wave vector (q2)
Run refinement
Refinement converges with the following Robs values for all reflections, main reflections, 1st
and 2nd order satellites: 10%, 3%, 13%, 51%
By second order satellites are considered reflections described by combination of q1 and q2:
q1 + q2 or –q1 + g2.
“Edit atoms”: add another two position modulation waves for all atoms
This mean one harmonic wave along the first wave vector combination (q1+q2) and one
harmonic wave along the second wave vector combination (-q1+q2)
Run refinement
Refinement converges with the following Robs values for all reflections, main reflections, 1st
and 2nd order satellites: 4%, 2%, 7%, 6%

11. Contour sections and ADP modulation

209 Jana2006 Cookbook, version October 2017

“Start Contour → New plot → Calculate new ones”
Select observed Fourier
For “Scope” define Ba1 as central atom
For map axes select x1, x4, x3, x2, x5 with limits as shown below

This will plot section x1-x4 but the shape of atomic domain will be function of x5
OK to calculate Fourier map
In Contour press “Atoms edit” and add Ba1

Press “Movie” to see the changes of this figure with x5

210 Jana2006 Cookbook, version October 2017

 Press “New plot” and calculate difference Fourier map for the same section

The difference maxima around Ba1 are very strong indicating ADP modulation

“Edit atoms”: set two ADP modulation waves for Ba1

Run refinement
Refinement converges with the following Robs values for all reflections, main reflections, 1st
and 2nd order satellites: 3%, 2%, 4%, 6% for 121 parameters

12. Refinement of inversiontwin

As the structure is non-centrosymmetric we cannot be sure that the proper absolute
configuration has been reached. To prove the correct configuration we shall refine so called
Flack’s parameter which means the volume fraction of twins related by the inversion centre.
“Start EditM50” to the page cell, define the twinning operation by inversion centre
Run “Parameters→ Twin fraction” and allow refinement of the twin ratio.
Run refinement
Refinement converges with the following Robs values for all reflections, main reflections, 1st
and 2nd order satellites: 2.1%, 1.5%, 2.6%, 5.5% for 122 parameters.
Run “Parameters→ Twin fraction” to see the refined volume fraction ~ 0.915. Which means
the absolute configuration is opposite to that we got when going from the centrosymmetric
to non-centrosymmetric description.
Make backup copy: “fresnoite_final”

13. Inverted barium position – false minimum

Barium is located in a special position x, x-1/2, 1/4. Application of a two-fold axis creates a
position -x, -x+1/2, 1/4, which becomes – after application of the three-dimensional part of
211 Jana2006 Cookbook, version October 2017
the X centering 0 0 1/2 1/2 1/2 – equivalent to the inverted position –x, -x+1/2, -1/4. Thus in
average structure position of barium can be inverted. In modulated structure inversion of
barium yields different modulation functions because of the 4th and 5th component of the
centering vector.

Refinement of fresnoite tends to be trapped in a false minimum because of very good fit of
main reflections and relatively weak satellites. Inversion of the position of barium helps to
leave the false minimum.

Start “Edit Atoms”, select Ba1

“Action → Transform selected atoms” and apply inversion -x1,-x2,-x3,-x4,-x5
Save the result, leave “Edit atoms”
Run refinement
Refinement converges with the following Robs values for all reflections, main reflections, 1st
and 2nd order satellites: 7.6%, 2.7%, 15.6%, 14.2%. This means that the structure can reach a
false minimum but also stable minimum.

212 Jana2006 Cookbook, version October 2017

Example 5.5.1: Melilite data
Data processing of a (3+2)-dimensional incommensurately modulated structure
with overlaps of satellites, using Crysalis.
Revised: 2 December 2016

Ca2MgSi2O7
Single crystal data measured with Oxford Diffraction four-cycle diffractometer
Input files: directory Melilite available at http://crysa.fzu.cz/jana/workshops/Melilite.zip

This example is based on the Crysalis 171.38.43. Crysalis is a program for controll and data
processing of Agilent diffractometers. It has very friendly licencing policy: everybody can install
Crysalis for processing data acquired with an Agilent instrument (which is also the case of
Melilite).

1. Getting the file

Start your browser
Connect to http://crysa.fzu.cz/jana/workshops/Melilite.zip and download the archive
Unpack the archive
It will create directory Melilite

2. Starting Crysalis
Start Crysalis
In a dialogue that follows locate the file run.par
Crysalis opens the experiment.
Use to decrease the image size

3. Determination of the unit cell

Start Lattice wizard using the button

[On the screen: the window of Lattice Wizard]
Start “Peak hunting” with default options
About 30000 peaks have been found
Start “Unit cell finding” with default options
The program finds a unit cell 7.84 7.84 5.02 90 90 90 and indexes about 70% of peaks

213 Jana2006 Cookbook, version October 2017

Start “Ewald Explorer” with default options
Use the buttons to see projections of the found peaks along the unit cell
axes.
Set projection along c*

The projection shows main reflections with radial distribution of spurious peaks. These peaks
are tails of very strong reflection and their radial districution allows quick conclusion that
these are not satellites.

Enlarge the peak projection using the right mouse button

Majority of satellite peaks can be described by two q-vectors: q1 = (α, α,0), q2 = (-α, α,0). The
214 Jana2006 Cookbook, version October 2017
 positive direction of a* and b* is indicated in the shrinked Ewald view by very small labels.

The peaks indicated by white circles are λ/2 reflections. The experiment was done with
classical Mo x-ray tube and graphite monochromator and with this equipment such
reflections may exist between very strong regular reflections.

The most important are weak reflections indicated by red circles. These reflection are
described by combination of q1 and q2, which proves that our structure is (3+2)-dimensional.
Without the combinations there would also be possibility of a (4+1)-dimensional crystal
twinned by 90° rotation around c*.

The combined satellites are also visible in the projections:

The projections also indicate that the combined satellites may be overlapped due to short
distance between them

In the next step, we will visualize the reciprocal plane 0,k,l and 1,k,l to see better the
combined satellites

4. Unwarping
Close the Ewald explorer
In Lattice wizard, press “Unwarping – precession images”
NEXT; NEXT
[On the screen: Step 3: Creation of a layer list”]
Press “Generate layers”
Type 1 for “max order”, 1.2 for “resolution”
Delete all layers except 1_k_l
NEXT
[On the screen: Step 4: Use background subtraction”]
Check “Use background subtraction”
NEXT
[On the screen: Step 5: Data corrections”]
Press “Edit user scale”
Define a user scale 40 (to see weak features)
Check “Don’t apply final convolution filter” (to see better separation of spots)
FINISH
Check the radio button “Layer” to see the progress of the unwarp procedure.
Wait until the layer is finished
Open the menu in the lower right cornes of Crysalis and select “Unwarp reciprocal lattice
grid”

215 Jana2006 Cookbook, version October 2017

 This is the resulting unwarped reciprocal plane (precession-like image):

The weak spots indicated by yellow rectangle are the λ/2 reflections. The ones indicated by
red circles are the combined satellites. Again, the possibility of overlaps between the
combined satellites is evident.

5. Setting up the (3+2)d structure, refinement of q-vectors

Start Lattice wizard and Ewald explorer
Press “Activate incommensurate peaks” and select “Add 2d modulation vectors (hklmn)”
Define two q-vectors: (0.3, 0.3, 0) and (-0.3, 0.3, 0) with maximum satellite index 1

216 Jana2006 Cookbook, version October 2017

 Using the sliders, modify the componets to fit with the projected peaks

Close Ewald explorer

In Lattice wizard, press “Incommensurates/Quasi--crystals”
Use default settings (with mmax = 1); OK
The refined unit cell + q-vectors index 92% of peaks. The procedure could index about 1800
combined satellites.

6. Combined satellites in Jana2006 indexing tool

Using Jana2006 indexing tool is not necessary for getting results in this example.
Open options of “Save information” and select “Save specific peak table only”
Save peak table as, e.g., forjana.tabbin
217 Jana2006 Cookbook, version October 2017
 Quit Lattice wizard
Start Jana2006 and open any structure
Start Tools → Graphic → Indexing
In the file manager of Jana2006, locate forjana.tabbin; double click it; OK
Select Mo KAlpha radiation; OK
[On the screen: Jana2006 indexing tool]

Press “View direction” and select c*; OK

Press “All to cell ON”

The resulting image projects all peaks positions into one unit cell. We can recognize q1 and
q2 satellites, combined satellites (the double-points in the middle of the unit cell axes) and
the λ/2 reflections in the middle of the plot.
Use to determine satellites coordinates
Use Refine cell to obtain refined q-vectors
Quit Jana2006

7. Data reduction
In the right panel of Crysalis, press “Start/Stop” and select “Data reduction with options”
218 Jana2006 Cookbook, version October 2017
[On the screen: Step1: Orientation matrix for data reduction]
Select “Other (reduction list)”
OK to confirm a message about non-existing reduction list
Press “Generate”
[On the screen: Proffit reduction list generation]
This tool generates list of reflections, which will be used for the data reduction.
Crysalis could also read an existing list of reflections (the “Load” button). This list can be
produced by Jana2006 and it is very useful for cases when the list cannot be generated by
Crysalis, e.g. for composite structures or for structures with X centering vectors. In this case,
however, we can generate the list by Crysalis.

Activate (3+2)d type of the list

Check “use cross q-vector reflections”

219 Jana2006 Cookbook, version October 2017

In the case of a structure with well separated spots, the generated list could be used for all
data.

In the case of Melilite, the combined (cross) satellites need a special treatment. The previous
attempts showed that these satellites are partially overlapped and that the best way how to
process them is making the data reduction with a large integration mask. With such mask,
the pairs of close combined satellites are integrated as a one spot and this can be taken into
the account in the Jana2006 refinement. However, this large mask is not suitable for main
reflections and for the q1/q2 satellites. For this reason, the combined satellites must be
processed separately.

Step 1 - data reduction for q1 and q2 satellites

Clear “use cross q-vector reflections”

OK; NEXT; NEXT

In the Step 3, clear “Follow significant sample wobbling”; NEXT; NEXT

220 Jana2006 Cookbook, version October 2017

In the Step 5, use outlier rejection for 4/m; NEXT
In the Step 6, press “Change output name” and define “q1_xor_q2” as an output file name
OK; FINISH
The program makes data reduction for main reflections, q1 satellites and q2 satellites, using
the default integration mask

In the left toolbar of Crysalis, press

Select the page “Data reduction output” and look at Rint of main reflections and satellites

The Rint of satellites is poor but we should consider that the program does not show
separately the observed satellites. Rint of observed satellites will be shown in the Reflection
report of Jana2006.
Select the page Graphs

The graph of frame scaling is almost flat, indicating sufficient absorption correction

221 Jana2006 Cookbook, version October 2017

Step 2 - data reduction for combined (cross) satellites

In the next step, we will process the combined satellites. We want to ignore the real position
of the cross satellites, use instead the position between the pairs of close satellites and large
integration mask to cover the complete pairs. For this, we have to disable refinement of
instrument parameters and profile analysis.
Quit the tool “Inspect data reduction results”
Start Lattice wizard
Use the “>” button at “Incommensurate/qusicrystal” and set q-vectors
q1 = ¼ ¼ 0
q2 = -¼ ¼ 0

Close Lattice wizard

In the right panel of Crysalis, press “Start/Stop” and select “Data reduction with options”
[On the screen: Step1: Orientation matrix for data reduction]
Select “Other (reduction list)”
Press “Generate”
In the line for q1, press “Edit mmax” and define mmax=0
In the line for q1, press “Edit mmax” and define mmax=0
Check “Use cross q-vector reflections”
Press “Edit max index” for the cross q-vector reflections and define mmax=1

OK; YES to confirm that we really want such a list

Out of Crysalis, go to the experimental directory and open in a text editor the file
run.dcredhklmn

222 Jana2006 Cookbook, version October 2017

The file only contains main reflections. This means the current version of Crysalis cannot
generate simply the list we need. However, we can use the full list + a custom extinction rule

CANCEL; CANCEL
In the main window of Crysalis - the left toolbar, press the button CMD
Type command “dc extinct hklmn 0 0 0 1 1 2 1”; ENTER
This defines that reflections with m+n odd will be skipped
In the right panel of Crysalis, press “Start/Stop” and select “Data reduction with options”
[On the screen: Step1: Orientation matrix for data reduction]
Select “Other (reduction list)”
Press “Generate”
In the line for q1, press “Edit mmax” and define mmax=1
In the line for q1, press “Edit mmax” and define mmax=1
Check “Use cross q-vector reflections”
Press “Edit max index” for the cross q-vector reflections and define mmax=1

At the bottom of the “Generate hkl list” window we can see that our custom extinction rule
is active. It can be checked by “Show rules” button
OK;
Crysalis displays warning about missing “nada/um q-vector refinement” but this is OK
because or q-vetors should not be refined
Out of Crysalis, go to the experimental directory and open in a text editor the file
run.dcredhklmn

223 Jana2006 Cookbook, version October 2017

 Now the file contains the combined satellites and main reflections. Main reflections can be
used for verification of the same scale for both data sets during their combining in
Jana2006, and discarded
(Back in Crysalis)
NEXT; NEXT
[On the screen: Step3: Basic algorithm parameters]
Press “Edit special pars”
Check “Override integration mask size”
Define size 1.75 (correct number can only be found by trials and errors)
Press Alt-E to display hidden options
Select “Skip model refinement” and “Skip profile fitting”

OK; OK; NEXT; NEXT;

For outlier rejection, select Lause symmetry “1”, Friedel pairs equivalent
NEXT;
OK; “Change output name” -> q1_and_q2
FINISH

8. Results of data reduction

After finishing the data reduction, start “Inspect data reduction” by pressing
The results for combined satellites are poor but this might be caused by their weakness. We
cannot make any conclusion without refinement of a structure model.

224 Jana2006 Cookbook, version October 2017

The input files for Jana2006 are in the directory /struct/tmp:
q1_xor_q2.hkl; q1_xor_q2.cif_od
q1_and_q2.hkl; q1_and_q2.cif_od

The example 5.5.2. where structure of melilite is solved uses different input files: 4d data for
q1, 4d data for q2 and 5d data for combined satellites. The reason is that we need to have
the user-independent results. Data created in this example are tested at the end of the
Example 5.5.2.

225 Jana2006 Cookbook, version October 2017

226 Jana2006 Cookbook, version October 2017
Example 5.5.2: Melilite
Solution of a (3+2)-dimensional incommensurately modulated structure with
overlaps of satellites.

Ca2MgSi2O7
Revised: 2 December 2016
Single crystal data measured with Oxford Diffraction four-cycle diffractometer

The structure has two q-vectors: q1=(α,α,0) and q2=(-α,α,0). Satellite reflections h,k,l,m,0 and
h,k,l,0,n are always well separated in the CCD frames. Combined satellites hklmn, m≠0, n≠0, are
sometimes separated and sometimes overlapped, depending on their geometry in particular
CDD frame (i.e. angle ψ). The combined satellites needed different data processing strategy
using large integration mask because we want to treat them like “always fully overlapped“.

Schematic view of the diffraction pattern in melilite. Main reflections are indicated by large
dark circles, small circles symbolize first-order satellites (grey for m x n = 0 and white for m x n =
±1 satellites, respectively).

Input files:
q1.hkl, q1.cif_od (main reflections + satellites described by q1)
q2.hkl, q2.cif_od (main reflections + satellites described by q2)
q1q2.hkl, q1q2.cif_od (main reflections + combined overlapped satellites)
Frame scaling, absorption correction: done with software of the diffractometer

1. Creating new jobname

Start Jana2006
“File → Structure → New” opens a file manager
Left pane: locate directory with input files
Right pane – textbox: type “melilite”
OK

2. Import Wizard
Select “Known diffractometer formats”; NEXT

227 Jana2006 Cookbook, version October 2017

Select “Oxford Diffraction – CCD”;
For “File name” select q1.hkl; NEXT
In “Define basic input file“ select “Input from cif_od file“; OK
Change the modulation vector to 0.278 0.278 0; NEXT
The used q-vector refinement software (CrysAlis) does not allow restriction that would keep
all components of the two q-vectors the same
[On the screen: Define the reference cell / split by twinning]
Change target dimension to 5
For 2nd modulation vector type -0.278 0.278 0; NEXT
The program reads 92323 reflections
For absorption correction select “None or done before importing”; NEXT
FINISH
[On the screen: Data repository]
Import new
Select “Known diffractometer formats”; NEXT
Select “Oxford Diffraction – CCD”;
For File name choose q2.hkl; NEXT
In “Define basic input file“ select “Input from cif_od file“; OK
Change the 1st modulation vector to -0.278 0.278 0; NEXT
Check “Import only satellites”;
New import of the same main reflections would bias their weighting scheme
Define transformation matrix (1 0 0 0 0 | 0 1 0 0 0 | 0 0 1 0 0 | 0 0 0 0 1 | 0 0 0 1 0); OK
This will convert hklm from the input file to hkl0n
NEXT
The program reads 92522 reflections, including discarded main reflections
For absorption correction select “None or done before importing”; NEXT
FINISH
[On the screen: Data repository]
Import new;
Select “Known diffractometer formats”; NEXT
Select “Oxford Diffraction – CCD”;
For File name choose q1q2.hkl; NEXT
In “Define basic input file“ select “Input from cif_od file“; OK
Change the 1st modulation vector to 0.278 0.278 0
Change the 2nd modulation vector to -0.278 0.278 0; NEXT
Check “Import only satellites”; NEXT
[On the screen: warning about changed unit cell parameters]
“Continue with the old reference cell parameters”; OK
The program reads 147857 reflections, including discarded main reflections
For absorption correction select “None or done before importing”; NEXT
FINISH
OK; YES to accept the data sets

228 Jana2006 Cookbook, version October 2017

3. Symmetry Wizard
Use all data blocks for the space group test

[On the screen: Tolerances for crystal system recognition]

For “Maximal ave(I/sig(I)) for extinctions type “20”; NEXT
Select tetragonal Laue symmetry 4/mmm
Select P centering
Select space group P-421m
At this point there is no difference between P-421m and P4212, the decision is expected
from the symmetry test of Superflip
Accept the superspace group in the original cell
For five-dimensional structures the wizard does not offer selection of superspace group. It
will be done later on manually.

4. Creating refinement reflection file

Follow steps of the reflection wizard, the data sets are on the same scale
The program discards 206 reflections, 96 observed, Rint=3.31%
In case that the same scale of the data sets would not be sure, a temporary structure should
be created with data sets imported together with main reflections. Then we could select
“different scales” and Jana would use main reflections for finding scaling factors. Then the
data set would be re-imported with the found scale factors and with main reflections only in
the first data set.

5. Structure solution wizard

Quit

6. Solution of average structure

Superflip for five-dimensional structure returns five-dimensional electron density map.
However, Jana2006 would not evaluate initial modulation functions from this map and we
would get only the average structure. For this reason solution will be done separately for
average structure
Start “Tools → Transformation → Go to average 3d structure”
For name use “aver”
Follow steps of the Reflection wizard
Continue with the new structure
Start “Run → Structure solution”
For Formula type “Ca Mg Si O”
Select “Superflip” and “Peaks from Jana2006”
Select “Use a specific random seed” and type “333” for random seed
With this option result of charge flipping will be the same in most computers and origin
position and special atomic position will be equivalent with this cookbook. For normal work
Random seed should not be fixed.
“Run solution”
229 Jana2006 Cookbook, version October 2017
 Superflip converges with R around 13% and confirms the selected space group
“Accept last solution”
Start EditM50; Go to the page “Composition”; for “Formula units” type “2”; press “Formula
from M40”age
It returns Ca2 Mg0 Si3 O7. The expected formula is Ca2 Mg Si2 O7, therefore one of silicon
atoms is probably magnesium.
Run Refinement
Refinement converges with Robs around 14%
Open “Dist” commands; clear “d(max) derived from atomic radii and typical distances”; run
“Dist”
Open listing of “Dist” and check Si-O distances.
Si2-O3 distances are unusually large (around 1.921 Å), that means Si2 should be changed to
magnesium
Using “Edit Atoms”, change atomic type of Si2 to Mg and its name to Mg1
Run refinement
Refinement converges with Robs around 14%
“Edit atoms”: change ADP of all atoms to harmonic
Run refinement
Refinement converges with Robs 3.7%
“Plot structure”

The structure looks normally but ADPs are unusually large

7. Back to the modulated structure

“File → Structure → Open”: open “melilite”
“EditM50, Composition page”: type formula Ca2 Mg Si2 O7, Z=2
Close Editm50, save changes
“File → Structure → Copy in”: open “aver” and import only M40

230 Jana2006 Cookbook, version October 2017

 Run refinement to confirm R value for the average structure:

8. Superspace symmetry
We will need to complete the translation part of superspace symmetry operators.
“EditM50 → Symmetry”: focus the textbox with the space group symbol and press TAB.
The program shows two symmetry operators where the translation part in x4 and x5 needs
to be completed:

and

Each translation part has two elements which can be either 0 or 1/2. Thus we have sixteen
combinations: 0 0 + 0 0, 0 0 + 1/2 0, 0 0 + 0 1/2, 0 0 + 1/2 1/2 … However, some of them are
forbidden because they lead to a non-consistent translation vector, which cannot ne used for
generation of the superspace group. Finally, we have only eight combinations. In the
following table we show number of extinct reflections (observed/unobserved) which
Jana2006 reports for each of these possibilities:

Translations in x4,x5 00 1/2 1/2

00 96/206 181/5369
1/2 1/2 181/5369 96/206
1/2 0 96/206 181/5369
0 1/2 181/5369 96/206

Only four possibilities lead to additional extinctions. They are all equivalent as we can judge
from the exactly same number of extinct reflections as well as from the fact that EditM50
does not offer creation of new refinement reflection file when we switch between them.
Another four possibilities without additional extinctions are equivalent, too.
231 Jana2006 Cookbook, version October 2017
 Thus we have two possibilities for the superspace group: with and without additional
extinctions. The choice is simple because the additional extinctions mostly concern
unobserved reflections and therefore the symmetry leading to additional extinctions is
confirmed by the experiment.

In order to be consistent with the published structure we will choose 0 0 for the first
symmetry operator and 1/2 1/2 for the second one.

Select 0 0 and 1/2 1/2 for the first and the second symmetry operator, respectively
Close EditM50, save the result
Create new refinement reflection file

9. Refinement of modulated structure – position modulation

Start “File → Structure → Save As” and make backup copy: “melilite_backup”
Start “Parameters → Options”
Press “Define wave vectors”
For the third wave define 1 1
This means modulation wave vector given by a sum of the q vectors
For the fourth wave define 1 -1
This means modulation wave vector given by a difference of the q vectors
OK; OK; YES to save changes
“Edit atoms”: set two position modulation waves for all atoms
This means one harmonic wave with wave vector q1 and one harmonic wave with wave
vector q2
In Refinement commands switch to window “Modulation”; for “Overlap option” select
“Closest reflections”; “max. satellite index” 1; “max. difference in degs” 0.5
With this settings pair of combined satellites will be treated as fully overlapped reflections.
Run refinement
Refinement converges after approx. 20 cycles with the following Robs values for all
reflections, main reflections, ±(1,0) and ±(1,1) satellites: 23%, 10%, 28%, 36%
“Edit atoms”: add another two position modulation waves for all atoms
This means one harmonic wave with wave vector q1+q2 and one harmonic wave with wave
vector -q1+q2
Run refinement
Refinement converges with the following Robs values for all reflections, main reflections,
±(1,0) and ±(1,1) satellites: 9%, 6%, 13%, 11%

10. Contour sections and ADP modulation

“Start Contour → New plot → Calculate new ones”
(If no old map exists from previous calculations this option is not active but Contour will use
it)
Select observed Fourier
For “Scope” define Ca1 as a central atom
For map axes select x1, x4, x3, x2, x5 with limits as shown below

232 Jana2006 Cookbook, version October 2017

This will calculate section x1-x4 through Ca1 as a function of x5
OK to calculate Fourier map
In Contour press “Atoms edit” and add Ca1

Press “Movie” to see the changes of this figure with x5. ESC to stop.

233 Jana2006 Cookbook, version October 2017

 Press “New plot” and calculate difference Fourier map for the same section

The difference maxima around Ca1 indicate ADP modulation

Calculate and plot corresponding difference Fourier sections for the other atoms
All atoms have strong difference maxima and minima
“Edit atoms”: set two ADP modulation waves for all atoms
Run refinement
Refinement converges with the following Robs values for all reflections, main reflections, 1st
and 2nd order satellites: 6.1%, 3.7%, 5.3%, 11.1% for 166 parameters
Start Contour and repeat plotting of the difference section through Ca1
The difference maxima are not so large but still not acceptable
“Edit atoms”: set additional two ADP modulation waves for all atoms
Run refinement
Refinement converges with the following Robs values for all reflections, main reflections, 1st
and 2nd order satellites: 4.8%, 3.1%, 4.5%, 8.1% for 238 parameters

Run Grapht and investigate distances, bond valence sums etc. All obtained graphs will
depend on “u” coordinate, which can be changed by “Movie” button.

11. Application of data prepared in Example 5.5.1

Make backup copy of your structure
“File → Reflection file → Import/modify the reflection file”
[On the screen: Data repository]
Delete all data sets
“Import New”
Follow steps of the import wizard to import the file “q1_xor_q2.hkl”, i.e. main reflections +
q1/q2 satellites. The files are located in the struct/tmp directory of the CrysAlis job, input
and target dimension is 5

234 Jana2006 Cookbook, version October 2017

As reference cell parameters, use the ones from previous refinement because they already
have corrected q-vector components

Do not close the Data repository.

“Import New”
Follow steps of the import wizard to import only satellites from the file “q1_and_q2.hkl”.

Use the old reference parameters - this causes that 0.25 atrufucial components of q-vectors
will not be used

235 Jana2006 Cookbook, version October 2017

Accept changes in data repository
Follows steps to create refinement reflection file (both data sets are on the same scale)
Run refinement

Refinement converges with the following Robs values for all reflections, main reflections, 1st
and 2nd order satellites: 4.4%, 2.3%, 4.1%, 7.9% for 238 parameters

236 Jana2006 Cookbook, version October 2017

Example 5.6: Iodomol
A molecular modulated structure from single crystal data – molecular
modulations and tests of positional vs occupational modulation
Revised: 1 October 2017

Single crystal data measured with a four-circle diffractometer and CCD detector.
Input files: iodomol.p4p, iodomol.ram
absorption correction: done

1. Creating new jobname

Start Jana2006
“File → Structure → New” opens a file manager
Left pane: locate directory with input files
Right pane: double-click Iodomol

2. Import Wizard
Select “known diffractometer formats”; NEXT
Select “Bruker CCD”;
Click the button “Browse” and double click “Iodomol.ram”; NEXT
[On the screen: complete/correct experimental parameters]
Leave all settings default; NEXT;
[On the screen: Define the reference cell/split by twinning]
Set target dimension to 4; NEXT; OK TO confirm the import of 52313 reflections
In the text window, leave all settings default; NEXT; Finish

3. Data Repository
The line in data repository window corresponds to one data set.
OK; YES to accept the data set
[On the screen: “Information about starting symmetry wizard”]
NEXT

4. Symmetry Wizard
[On the screen: “Tolerances for crystal system recognition”.]
Symmetry wizard can be started separately by “File → Reflection file → Make space group
test”.
Leave all settings of the wizard default; NEXT
[On the screen: “Select supercell”.]
Confirm “continue with the basic unit cell”; NEXT
[On the screen: Select Laue symmetry – the Rint values are quite elevated]
Confirm monoclinic Laue symmetry; NEXT
[On the screen: Select cell centering]
Confirm primitive unit cell; NEXT
[On the screen: Select space group]
Confirm P21/m; NEXT
[On the screen: Select superspace group]
Select P21/m(a0g)0s – out of 1038 potentially absent reflection none is observed, this space
237 Jana2006 Cookbook, version October 2017
 group is thus the most likely.
[On the screen: Final step of the superspace test]
Accept the superspace group transformed into the original cell: P21/m(a0g)0s; FINISH

5. Creating refinement reflection file

We have finished creation of the file M95 (reflections converted from diffractometer format
to Jana format) and M50 (cell parameters, symmetry). At this step, we will create the file
M90 containing data set merged by symmetry, with forbidden reflections omitted.
NEXT to confirm threshold 3sigma;
OK; OK to confirm information about imported reflections
[On the screen: settings for merging of reflections]
Leave all settings default; NEXT
The Rint for observed reflections will be 19.57%
OK
Accept the data block and calculate coverage; FINISH; OK
[On the screen: Information that structure solution will start]

6. Structure Solution Wizard

[On the screen: window of Structure solution]
In “Formula” textbox type list of chemical elements: C O H I P
Check the check box “Use a specific random seed”
This option is here only to ensure the reproducibility of the calculation. In normal practice, it
is not necessary.
Keep all other settings default
The default solution program in Jana is Superflip. If you have other software available (SIR,
SHELXT), you may link it to Jana in Tools/Programs
Click “Run solution”
Superflip converges within a few seconds. It confirms the symmetry P21/m(a0g)0s.

238 Jana2006 Cookbook, version October 2017

 This is symmetry found
by Superflip in the
electron density map

This is fit of symmetry

elements supplied by
user (small numbers =
good fit)

Close the viewer of the listing, wait until the structure solution is interpreted
Click “Draw structure
Choose the radio button “Draw average structure”
If you have followed this tutorial exactly, you should see a picture identical to the following
(to obtain the picture in Diamond, click the icon “Get molecules directly” - )

If the structure looks slightly different (e.g. shifted by ½ along a or b), you may still continue
with the tutorial, but the details (atom names, atom positions) may be different.
Alternatively you may rerun Superflip until you obtain the solution in the correct setting.

239 Jana2006 Cookbook, version October 2017

 The structure looks correctly interpreted and complete. Only atoms C1 and C2 should be
oxygen atoms.

Close the structure viewer and exit the solution wizard by clicking “Accept last solution”

7. Checking the structure

First of all we will correct the wrong atomic types.
Double click the icon “Edit atoms”
Double click atom C1
Type “O*” in the text box “Name”; Press TAB
The name changes to the first free number for oxygen, O1 in this case
Change the atomic type to O;
Repeat the procedure for atom C2
OK; Yes to confirm the changes and return to the main window.

We will now check the modulation functions for all atoms

Double click the icon “Contour”
Click the button “New plot”
Select the radio button “Show de Wolff’s sections”
Set the scope of the sections: primary scope to 3, secondary to 0.5
Click the button “Define atoms for de Wolff’s sections
In the text box “Atoms” type * (a single star) and click the button “Add”; OK
OK
The program calculates the Fourier maps for modulation functions of all atoms (x, y and z
direction for each atom). It may take some time.
Go through the de Wolff’s sections for all atoms (using the buttons “W-“ and “W+” or by
selecting from the list and check, if the modulation functions suggested by Jana2006 follow
the domains in the Fourier maps. This is an example of the x2-x4 section for I1:

240 Jana2006 Cookbook, version October 2017

 You should conclude that all atoms have been interpreted correctly and the modulation
functions are a good starting point for the refinement. The only exception is the x1
modulation function of the atom C11.This could be corrected manually (invert the signs, shift
the average position), but the refinement is robust enough to correct this single problem
automatically.

Quit Contour and return to the main window.

8. Initial structure refinement

Important: Make sure you copy the job to a new job name to preserve the original and to
have the chance to return to this point.

Go to “File->Structure->Save as
In the lower right text box type “Iodomol-ref”; OK
Yes to continue with the new structure
Right-click the icon Refine
Set the number of refinement cycles to 100:

OK; Yes + Start

The refinement converges to the following R-values:

9. Refinement of ADP

So far the displacement parameters are refined as isotropic. We will now set them for
anisotropic refinement:

Open “Edit atoms”

Select all atoms
Double click any of the selected atoms
Select the radio button “harmonic (anisotropic)”
OK; OK; Yes to quit “Edit atoms”
Double click Refine to refine anisotropic displacement parameters.

The refinement converges to the following R-values:

241 Jana2006 Cookbook, version October 2017

 Sometimes the displacement parameters themselves are modulated. This may be important
especially for the heavy atoms. We can try to add one harmonic modulation wave for the
iodine atoms:

Open “Edit atoms”

Select the atoms I1 and I2
Double click any of the selected atoms
increase the number of ADP harmonic waves from 0 to 1”
OK; OK; Yes to quit “Edit atoms”
Double click Refine.

The refinement converges to the following R-values:

The R-value on the satellite reflections decreased by almost 2%, the modulation of ADPs
seems significant.

10. Defining phenyl rings as molecules

Save the structure as a new structure under the name iodomol-mol and continue with the
new structure;
Draw the average structure and check the geometry. Plot the structure with thermal
ellipsoids. The molecule looks like this:

242 Jana2006 Cookbook, version October 2017

 Ring 3

Ring 2

Ring 1

The result is not a disaster, but the displacement parameters are uneven. The data quality
may not be sufficient to refine all atoms independently. We can define the three phenyl rings
as three positions of the same molecule.

Note the atom names for the free phenyl rings. Marking the phenyl rings as in the image
above, you should get the following lists:
Ring 1: C3, C6, C9, C13, C18, C12
Ring 2: C5, C15, C20, C22, C10, C8
Ring 3: C16, C17, C19, C21, C14, C11
Close the structure viewer.
Go to Parameter -> Molecules -> New molecule;
Type “phenyl” in the text box “Name of the molecule”; Next
Select atoms C3, C6, C9, C13, C18, C12 from the list; OK;
Choose gravity center as the molecular reference point; Next;
Note that the occupancy of this position is set to 0.5, because Ring 3 is located on the mirror
plane.
Click “Apply + Next position”;

Add the next position by identifying the model atoms with the atoms of the next phenyl
ring. Fill the table according to the following picture:

243 Jana2006 Cookbook, version October 2017

Click “Show coinciding atoms”. You will obtain six coinciding atoms between the model
molecule and the new position:

Click “Apply + Next position”

Define the third position of the molecule:

Check again that all six atoms of the model molecule coincide with the actual atoms in the
atomic part of the model.
Click “Apply + End”

Now we defined all three phenyl rings as instances of a single molecule. The positions of
atoms in the molecule will still be refined, but one molecule will be placed in three positions
in the crystal. We also have to redefine the modulation parameters – remove the

244 Jana2006 Cookbook, version October 2017

modulation from the atoms and add modulation of the molecular parameters – position and
rotation.

Open “Edit atoms”

Select the six atoms shown in blue – these are the atoms belonging to the molecular part
Set the number of positional modulation waves to 0
Reset the displacement parameters to isotropic
OK; OK; Yes;
Go to Parameters -> Molecules -> Edit;
For each position increase the number of positional modulation waves to 1:

OK; Yes;

Double click Refine to refine the new model.

The refinement converges to the following values:

Note that the R-values are only marginally higher than in the previous refinement, but the
number of parameters decreased by a factor of two, from 344 to 176.

For future convenience we will now rename the atoms in the molecule sequentially as
C1…C6:

245 Jana2006 Cookbook, version October 2017

 Open “Edit atoms”
Select all atoms
Click Action -> Rename selected atoms manually
Rename atom C4 to C8 (to avoid name conflict with the newly named C4) and rename the
other atoms according to this image:

OK; OK; Yes

11. Alternative model – occupational modulation by crenel functions

Note: the following part includes some slightly more advanced techniques, and it turns
out the alternative model is not better than the previous one in this particular case. It is a
very useful exercise though. You are encouraged to follow it, but if you are short of time
or need to focus on something else, you may skip this part and continue with the next
section (12. – viewing the structure).

If you check carefully the de Wolff’s sections of some of the carbon atoms, you will see that
it seems more appropriate to describe the modulation not as a harmonic oscillation of the
atom, but as a discontinuous switching between two positions. A good example is the x
coordinate of atom C2b or C3b.

In the following section we will set up this modulation for the phenyl rings to check, if this
model may be more appropriate than the harmonic model.
Setting up the crenel occupational modulation is in general more difficult than the harmonic
modulation, but often the model fits better and makes more sense.

General idea: Rings 2 and 3 need to be split to two copies, each will be present only in one
half of the complete domain along x4 and the domains will be complementary along x4.
Ring 1 lies on a mirror plane and it does not need to be split. If its occupation domain is

246 Jana2006 Cookbook, version October 2017

restricted to 0.5, the second half will be generated automatically by symmetry, because the
mirror plane is associated with a one-half shift of modulation phase.

First save the structure as iodomol-crenel and continue with the new structure
Check the de Wolff’s sections for the atoms of the ring 3 (position 3 of the phenyl
molecule):
Go through the de Wolff’s sections of atoms C1a, C2a, C3a, C4a, C5a, C6a and try to make a
conclusion about the behavior of the ring as it is modulated. Note the x4 coordinate of the
largest deviations of the atoms from the average position.
Conclusion: the largest modulation is along the y (x2) coordinate. The smallest modulation is
on the C1c atom, which is bonded to the phosphorus. C4c has also a relatively small
modulation amplitude compared to the rest of the atoms. This indicates that the phenyl ring
is pivoting around the axis C1c-C4c. This is also confirmed by the shape of the modulation
functions of C2c, C3c, C5c and C6c. The first two have very similar modulation functions and
it is just the opposite of the latter two:

The occupancy domain should have a length of 0.5 and the center approximately 0.4. To
generate a perfect model, we should also rotate the molecule from its average position.
However, this is difficult to do, as the rotation angles of the molecule are, in general, not
simply related to the crystallographic axes. We will therefore leave the molecule as it is and
let its position refine.

Close Contour
Go to “Parameters -> Molecules -> Edit”
Set number of positional modulation waves for the Position 1 to 0
Check the check box “Use: crenel”
Set the parameter aimol from 0.5 to 1
The previous step is done because now,due to the shortening of the domain, the molecule
does not sit on a special position anymore, and the occupancy need not be reduced.
Click “Edit modulation parameters -> Occupancy”
Set delta to 0.5 and x40 to 0.4
Check the checkbox next to x40 to allow its refinement
OK; OK; Yes to confirm changes
Go back to the de Wolff’s sections and check that the occupation domain of all atoms of the
Ring 1 reduced to one half of the full length.

Make the analysis of the modulation functions of the atoms from Ring 2 and Ring 3. Can you
estimate the modulation characteristics of these rings similarly to Ring 1?
247 Jana2006 Cookbook, version October 2017
Conclusion for Ring 2: largest modulation along x2, all atoms have a similar modulation
phase (maxima and minima at the similar x4). Minima for x4 between -0.05 and 0.2,
average around 0.07. Modulation amplitude along x1 about 0.01 in fractional coordinates.
Conclusion – the main feature of the modulation of Ring 2 is the shifting as a rigid body
along x2
Conclusion for Ring 3: the largest modulation is along x1, all atoms have approximately the
same amplitude of the function along x1 and also the minimum and maximum of the
modulation function of all atoms is located at about the same x4 – minimum at ~0.15,
maximum at ~0.65.The amplitude of the modulation along x1 is about 0.04 in fractional
coordinates. Conclusion – the main feature of the modulation of Ring 3 is the shifting as a
rigid body along x1.
Rings 2 and 3 are not on a special position, and therefore we first need to split each of them
to two independent molecules

Go to “Parameters -> Molecules -> Expand”

Select positions Phenyl#2 and Phenyl#3; OK
Click “Next matrix”
Leave the unit matrix; OK; Yes
Two new positions were created. Position Phenyl#4 is identical to Phenyl#2 and Phenyl#5 to
Phenyl#3. We will now modify their parameters.
Go to “Parameters -> Molecules -> Edit”
For position #2 do the following:
- set number of positional waves to 0
- check the check box “Use: crenel”
- click “Edit modulation parameters -> Occupancy”
- set delta to 0.5 and x40 to 0.07 (based on your previous analysis)
- still in the same form, check the check box next to the parameter x40, so that this
parameter gets refined; OK
- change the parameter “ytrans” to 0.117 (i.e. shift by -0.01 from the original position)

Repeat the procedure for positions #3, #4 and #5 with these settings:
#3: x40=0.15, delta 0.5, xtrans -0.345 (i.e. shift by -0.04)
#4: x40=0.57, delta=0.5, ytrans=0.137
#5: x40=0.65, delta 0.5, xtrans -0.265
Do not to forget to reset the number of positional modulations to 0 for all positions!
OK; Yes to confirm changes

Check the de Wolff’s sections again to check that the new parameters indeed generated
crenel-like modulations and shifted the atoms to approximately the right positions. You
should see images similar to this (example for atom C2e, coordinate x1):

248 Jana2006 Cookbook, version October 2017

As a last step before the refinement we will restrict the centers of the occupational domains
of the related molecules so that they always remain complementary:
Quit Contour
Right-click the icon Refine
Go to the tab “Various”, click the button “Equations”
In the equation text box type “t40[Phenyl#4]=0.5+ t40[Phenyl#2]”
Click “Add”
In the equation text box type “t40[Phenyl#5]=0.5+ t40[Phenyl#3]”
Click “Add”
OK; OK; Yes – do NOT start the refinement yet!
It is wise to make another backup now just for the case something goes wrong
After you have made a backup through “File -> Structure -> Save as”, you can start the
refinement:
Double click the icon Refine to start the refinement.

The refinement converges to the following values:

The R-value on the satellites is higher than in the previous refinement, but we have not yet
introduced positional modulation of the newly generated positions.

Go to “Parameters -> Molecules -> Edit”

For all molecular positions set the number of positional modulation waves to 1 and change
the type of modulation functions to “Legendre polynomials in crenel interval”
OK; Yes to confirm changes
Double click Refine to start refinement.

249 Jana2006 Cookbook, version October 2017

 The refinement may be instable. You can improve the convergence by temporarily
decreasing the damping factor to e.g. 0.2.

The refinement converges to the following values:

The result is slightly better than the previous refinements with harmonic modulation.
However, the number of parameters is much larger. Checking the de Wolff’s sections shows
that the refined positional modulations are not very realistic, the amplitudes are large and
the shapes tend to converge to the harmonic shape. We can thus conclude that in this
particular case the harmonic modulation is probably a better model.

12. Adding of hydrogen atoms

We will return to the harmonic modulation model:

Open the structure iodomol-mol with the harmonic modulation of the molecules.
Double click “Edit atoms”
Select all carbon atoms
Click “Action -> Adding of hydrogen atoms”
Leave all settings default; OK
Ten hydrogen atoms have been added automatically: four in the atomic parts and six in the
molecular part.
OK; Yes to confirm changes
View the average structure to check, if the hydrogen atoms have been added correctly.
Hydrogen atoms have been added on all carbons of the phenyl molecule – Jana does not
recognize bonds between the molecular and atomic part. The atom H1C1 should be deleted.
Close the structure viewer
Double click “Edit atoms”
Right-click atom H1C1; Delete selected atoms; Yes
OK; Yes to confirm changes
Double click Refine to refine the structure
The refinement converges to the following values

250 Jana2006 Cookbook, version October 2017

13. Viewing and interpreting the structure

We have so far not looked at the modulated structure in detail. We will now try to interpret
the modulation.

Traditionally the modulation functions are interpreted with the help of plots of distances,
angles, bond-valence sums etc. as a function of the modulation phase. This may be difficult
for molecular crystals, where the modulation is usually quite complex. It is more useful to
view the structure either in a supercell or as an animation of the modulation.

Plotting the supercell

Open the structure iodomol-mol with the harmonic modulation of the molecules.
Double click the icon “Plot structure”
Select “Draw approximant structure”
Set the range of x from 0 to 5 and for y from 0 to 6
5x6 supercell closely corresponds to the reciprocal values of the modulation vector
components of 0.19 and 0.15. We should see a good representation of various phases of the
modulation.
Click “Draw+continue”
In Diamond fill the unit cell and draw bonds (icons and ).
If Diamond is too slow, you can speed it up by switching off the rendering – icon
The best view of the structure is along a. You can see how the rows of the hydrogen iodide
molecule (only iodine present) form a wave, followed by the oxygen, to which the hydrogen
iodide molecule forms a hydrogen bond. Therefore the distance iodine-oxygen remains quite
constant. Also the large changes of the torsion angle of Ring 3 are visible throughout the
supercell. However, it is still difficult to make sense of the interactions between the molecule
and the overall picture of the modulation.

Looking at the movie of the modulation

The program MoleCoolQt can read and display modulated structure refined by Jana.
Unfortunately, so far it cannot read modulated molecules. To see the movie, we first need to
transform the molecular model to purely atomic model.
First save the structure as iodomol-display and continue with the new structure
Open “Edit atoms”
Select all carbons and hydrogens from the phenyl molecule (in blue)
Click “Action -> Atoms from molecule to atomic part
OK; Yes
Start the program MoleCoolQt
Go to “File -> Open…”, locate the file iodomol-display.m50 and open it
You may also drag an icon of M50 displayed in your file manager on the icon of MoleCoolQt
Keep the selection “Load as modulated structure”; Open
Switch off the atom labels – icon in the top toolbar
You should obtain this image:

251 Jana2006 Cookbook, version October 2017

Press the key t on the keyboard

Modulation movie starts. It shows the molecule as a function of modulation phase. Keep in
mind that this is NOT a movie of the actual molecular motion, although it is tempting to
interpret it like this! These are snapshots of the unit cell configurations at various modulated
phases, which are distributed across the whole crystal.

You may use the options in the menu “Pack” to fill larger part of the structure to see the
intermolecular interactions and their changes with the modulation phase.

252 Jana2006 Cookbook, version October 2017

Example 5.7: CaCNi
A simple molecular modulated structure from single crystal data with crenel
functions
Revised: 1 October 2017

Single crystal data measured with KUMA four-circle diffractometer and point detector.
Reference: Petříček, Dušek and Černák (2005), Acta Cryst. B61, 280-286
Input files: cacni.hkl
absorption correction: done

1. Creating new jobname

Start Jana2006
“File → Structure → New” opens a file manager
Left pane: locate directory with input files
Right pane: double-click CaCNi

2. Import Wizard
Select “reflection file corrected for LP and abrorption”; NEXT
Select “General file on I”;
In the text box “Input format” Type “*” (single star, without quotes); NEXT
[On the screen: complete/correct experimental parameters]
Fill unit cell parameters: 8.4867 15.9511 7.6038 90 90.48 90
Type the modulation vector on the “1st modulation vector” box: 0.4365 0 0.1307; NEXT;
In the text window, leave all settings default; NEXT; OK; Finish

3. Data Repository
The line in data repository window corresponds to one data set.
OK; YES to accept the data set
[On the screen: “Information about starting symmetry wizard”]
NEXT

4. Symmetry Wizard
[On the screen: “Tolerances for crystal system recognition”.]
Symmetry wizard can be started separately by “File → Reflection file → Make space group
test”.
Leave all settings of the wizard default; NEXT
[On the screen: “Select supercell”.]
Confirm “continue with the basic unit cell”; NEXT
[On the screen: Select Laue symmetry]
Confirm monoclinic Laue symmetry; NEXT
[On the screen: Select cell centering]
The data were measured with C-centering assumed. Confirm C centered unit cell; NEXT
[On the screen: Select space group]
Confirm C2/c; NEXT
[On the screen: Select superspace group]
Confirm C2/c(a0g)0s suggested by Jana based on systematic absences
253 Jana2006 Cookbook, version October 2017
 [On the screen: Final step of the superspace test]
Accept the superspace group transformed into the original cell: C2/c(a0g)0s; FINISH

5. Creating refinement reflection file

We have finished creation of the file M95 (reflections converted from diffractometer format
to Jana format) and M50 (cell parameters, symmetry). At this step, we will create the file
M90 containing data set merged by symmetry, with forbidden reflections omitted.
NEXT to confirm threshold 3sigma;
OK; OK to confirm information about imported reflections
[On the screen: settings for merging of reflections]
Leave all settings default; NEXT
The Rint for observed reflections will be 2.82%
OK
Accept the data block and calculate coverage; FINISH
[On the screen: Information that structure solution will start]
OK

6. Structure Solution Wizard

[On the screen: window of Structure solution]
In “Formula” textbox type list of chemical elements: Ni Cd C N H
Check the check box “Use a specific random seed”
This option is here only to ensure the reproducibility of the calculation. In normal practice it
is not necessary.
Keep all other settings default
The default solution program in Jana is Superflip. If you have other software available (SIR,
SHELXT), you may link it to Jana in Tools/Programs
Click “Run solution”
Superflip converges within a few seconds. It confirms the symmetry C2/c(a0g)0s.

254 Jana2006 Cookbook, version October 2017

 This is symmetry found
by Superflip in the
electron density map

This is fit of symmetry

elements supplied by
user (small numbers =
good fit)

Click “Draw structure” (select Draw average structure; Draw+continue)

If you have followed this tutorial exactly, you should see a picture identical to the following,
after filling the unit cell with atoms and viewing it down the c-axis:

If the structure looks slightly different (e.g. shifted by ½ along a or b), you may still continue
with the tutorial, but the details (atom names, atom positions) may be different.
Alternatively you may rerun Superflip until you obtain the solution in the correct setting.

255 Jana2006 Cookbook, version October 2017

 Close the structure viewer and exit the solution wizard by clicking “Accept last solution”

7. Checking the structure

All atoms in the structure are located in the planes a-b at z=0.25, or in the plane b-c, x=0.5.
We shall view the structure in these planes to understand it better.

Double click the icon Contour in the main window of Jana:

Click “New plot”; OK;
You see now the Fourier map of the structure. A few settings need to be modified to see the
relevant features:
Click t-map ON; OK
Click the button “Go to” and set x3 to 0.25
Click the button “Contours” and set positive contours to 3; OK
Click on each atom in the view and check the check boxes to draw the position of the atoms
(max. distance up to 1 Å)
You should obtain an image similar to this (inverted colors)

Obviously, Jana placed a single carbon atom in place of every cyano group. There is also a
split Cd atom of which only one position is marked. However, this picture is just one of many
unit cells in the modulated structure. To see the complete picture, we will run a movie of all
configurations:
Click the button “Movie”, set the start and end of x3 to 0.25, set delay time to 0.5 s and run
the movie by clicking OK. The movie can be interrupted by pressing Esc. Watch the movie
carefully and try to understand, what happens in the structure.

256 Jana2006 Cookbook, version October 2017

 These are the conclusions we can make from the movie:

- Atoms C2 and C3 should be deleted and replaced by properly placed C and N atoms of
the cyano group
- Atom Cd is modelled by a continuous modulation function, but in fact it is jumping
between two positions and it is better modelled with a split position.

We shall now make these modifications by first deleting atoms Cd1, C2 and C3 and then
placing them in the correct positions:

Quit Contour by clicking the button Quit

Double click the icon “Edit atoms”
Select Cd1, C2 and C3 (hold Ctrl while clicking the atoms), right click and select Delete
selected atoms; Confirm deleting by “Yes to all”. OK to close the form, Yes to confirm
changes.

Return to Contour and make the same plot as before.

Click t-map ON; OK
Click the button “Go to” and set x3 to 0.25 (as before) and t to 0.2.
Set positive contours to 3
Click the lower of the two split positions of Cd according to the following picture:

In the form that opens, click “Include the selected point as a new atoms in M40”
Fill the next form according to this picture (i.e. atoms Cd1, atomic type Cd):

257 Jana2006 Cookbook, version October 2017

OK;
Repeat the procedure with the second split Cd position and with atoms C2, C3, N2 and N3. You
should add six atoms according to the following scheme:

N3

C3

C2
Cd2
N2
Cd1

Please make sure to follow the scheme, so that the atomic names further in the tutorial
correspond to your structure model.

If you now click on the Cd atoms and let Jana show the position, you will probably see
something like this:

The Cd atom sits on the special position (two-fold axis) and the clicked position was slightly
away. We will correct this offset later.

Click the button Quit and confirm adding new atoms in the structure model.

Now we will similarly check the other section of the structure, the plane b-c at x=0.5:

Double click the icon Contour;

Click the button “New plot”
On the left check the radio button Calculate new ones; OK

258 Jana2006 Cookbook, version October 2017

In the subsequent form open the tab “Scope”, uncheck the box “Use default map orientation”
and define the b-c plane as the primary plane of the Fourier map according to the following
scheme:

Select the radio button “Whole cell; OK

Click the button t-map ON; OK
Set Contour level to 3 (button Contours)
Click the button “Go to” and set x1 to 0.5, t = 0
Click atoms in the view and draw their position. You should obtain this result:

Run movie (button movie, set x1 range to 0.5-0.5, delay time to 0.5) and investigate the
behavior of the modulation in this plane. Do we need to modify the structure model?

Your conclusion should be that the model is essentially OK. Atom C4 is modulated between two
positions and is always present close to the neighboring Cd atoms. The Cd atom oscillates
between four-fold and two-fold coordination (in this plane!). Atom C4 should be nitrogen
(ammonium). Atom C1 should also be a nitrogen atom. We can now make the final corrections
to the initial model:

Click “Quit” to quit Contour

Double click the icon “Edit atoms”
Double click atom Cd1
on the tab “Edit” change the x-coordinate to 0.5
Click the button Show/reset site occupancy and update the site occupancy to 0.5; Yes; OK
Repeat the procedure with the atom Cd2
Double click the atom C4
Type “N*” in the text box “Name”; Press TAB

The name changes to the first free number for nitrogen, N1 in this case

259 Jana2006 Cookbook, version October 2017

Change the type to N;
Repeat the procedure for the C1 atom. It will become N4
OK; OK; Yes to confirm the changes and return to the main window.

Now we have checked and updated the basic structure – atomic types and positions. Now we
can check the modulation functions of individual atoms. Jana tries to assign harmonic
modulation functions to each atom after the solution.

8. Setting up modulation functions

We will now check the modulation function for all atoms and modify them where necessary.
Double click the icon “Contour”
Click the button “New plot”
Select the radio button “Show de Wolff’s sections”
Set the scope of the sections: primary scope to 3, secondary to 0.5
Click the button “Define atoms for de Wolff’s sections
In the text box “Atoms” type * and click the button “Add”; OK
OK
The program calculates the Fourier maps for modulation functions of all atoms (x, y and z
direction for each atom). It may take some time.
Go through the de Wolff’s sections for all atoms (using the buttons “W-“ and “W+” or by
selecting from the list and make notes about the type of modulation for each atom and for
special function note their centers (denoted x40, it is the x4 coordinate of the center of the
domain) and widths (length of the domain along x4, denoted Δ).

You should obtain the following conclusions:

- Ni1. Saw tooth function: x40= 0.75 (exactly, by symmetry), Δ=1
- N4: harmonic, existing function is a good first approximation
- N1: crenel function, x40= 0.75 (this time not fixed by symmetry, the atom is not on special
position), Δ=0.5
- Cd1: crenel function, x40= 0.25 (exactly, by symmetry), Δ=0.5
- Cd2: crenel function, x40= 0.75 (exactly, by symmetry), Δ=0.5; Cd atoms have
complementary Crenel intervals
- N2 atom: needs to be split into two positions, N2 and N2b. N2 has a crenel function with
x40= 0.59, Δ=0.5; for N2b x40= 0.09 Δ=0.5. Splitting is the most prominent in the x-direction.
x(N2)=0.21, x(N2b)=0.29
- C2 atom: needs to be split into two positions, C2 and C2b. C2 has a crenel function with
x40= 0.57, Δ=0.5; for C2b x40= 0.07 Δ=0.5. Splitting is the most prominent in the x-direction.
x(C2)=0.12, x(C2b)=0.19
- C3: continuous function, can be harmonic
- N3: continuous function, can be harmonic

We will now set up these functions for individual atoms

Quit Contour
Double click “Edit atoms”
Double click atom Ni1
Click the check box “Use crenel”
In the form that opens set delta to 1, x40 to 0.75; OK
260 Jana2006 Cookbook, version October 2017
Select the radio button “Legendre polynomials in crenel interval”
Note: we defined the function as crenel, although it looks like a saw tooth. Legendre
polynomials contain a linearly sloping term, and a crenel function can turn to a saw tooth
during the refinement.
Repeat the procedure for N1, Cd1, Cd2, set the parameters according to the list above
Right click atom N2, select “Split atomic position”
Set the atom name to N2b; Uncheck the option “Generate restrict commands for REFINE”;
Apply
Double click atom N2 and set its Crenel function according to the list above. On the “Edit”
tab set its x-coordinate to 0.21. Repeat for N2b, set parameters according to the list above.
Repeat the whole splitting procedure, setting of Crenels and change of x coordinates for the
atom C2.

As a last step we will release the width of the crenel domain of Cd1 for later refinement.
Only this parameter needs to be explicitly released, other domain widths and centers will be
restricted before the refinement – see below.
Double click the atom Cd1
On Card “Edit” click “Occupancy” next to the text “Edit modulation parameters”
Check the check box next to “delta”
OK; OK; OK; Yes to quit “Edit Atoms”

Now we will check that the functions were correctly set:

Double click Contour
Calculate new de Wolff’s sections; use the option “Update existing sections” in the form
Go through the de Wolff’s sections again and check that the functions are properly set.
Correct the mistakes you might have done.
These are the sections for Ni1:

The difficult part of the work is over. We have set up the structure model, we have
understood the structure in quite a detail and now we can start the refinement from a good,
elaborated structure model.
Quit Contour and return to the main window.

261 Jana2006 Cookbook, version October 2017

9. Initial structure refinement

Important: what you have done up to now is very precious. Make sure you copy the job to
a new job name to preserve the original and to have the chance to return to this point.

Go to “File->Structure->Save as
In the lower tight text box type “CaCNi-ref”; OK
Yes to continue with the new structure
Plot the average structure. Use “Draw and Continue”, keep the structure view open and
return to Jana2006
Right-click the icon Refine
Set the number of refinement cycles to 100 and damping factor to 0.5:

The last step we need to make before the refinement is setting up restrictions on the
modulation parameters. The logic of the structure requires that the groups Cd1+N2b+C2b
have the same “domain of existence”, i.e. are present together in the structure. The group
Cd2+N2+C2 should also be present together and the two groups are mutually exclusive.
Furthermore, no vacancy is expected, i.e. either Cd1 or Cd2 is always present. Thus, the
centers of the crenel domains within each triplet should be equal and between the triplet
shifted by 0.5. The length of the domains is equal within each triplet, and the domain
lengths of one triplet are complementary to 1 with the domain lengths of the second triplet.
We need to define these relationships by means of equations between the atomic
parameters:

Go to card “Various”; Click the button “Equations”

In the text box “Equation” type delta[N2b]=delta[Cd1]
Click the button Add to add the equation
Add further equations restricting the domain lengths (parameter delta) and domain centers
(parameter t40).
The final list of equations will look like this

Note: There is no equation on t40 of Cd2, because this parameter is fixed by symmetry.
262 Jana2006 Cookbook, version October 2017
 OK; OK; Yes + Start
The refinement converges to the following R-values:

It is a promising start. The high R-values for the 2nd and 3rd-order satellites are not
surprising, we have not added higher-order positional modulation yet.

10. Add harmonic waves

Satellites up to third order are observed. We may add up to three harmonic waves, but first
we will add only one waves for each light atom and two waves for the heavy atoms.
Open “Edit atoms”
Double click Ni1
Increase the number of positional waves to 2
Repeat for other atoms, use 2 waves also for Cd1 and Cd2 and 1 wave for all other atoms.
Close Edit atoms and run refinement again.
The refinement converges to the following R-values:

Recalculate de Wolff’s sections and check, if the newly refined modulation functions are still
follow the traces in the Fourier maps.

The sections look OK, the modulation functions are appropriately describing the electron
density. The only exception may be the Cd1 and Cd2 atoms, where the functions are too
curved towards the end of the occupancy domain:

263 Jana2006 Cookbook, version October 2017

 The number of waves may thus be decreased to 1.However, testing leads to a significant
increase of the R-values and we therefore keep two modulation waves on Cd1 and Cd2 at
least for now.

Checking the shapes of the modulation waves is an important check to be made after every
significant change of the structure model to avoid divergence of the model to unrealistic
parameters.

11. Refinement of ADP

So far the displacement parameters are refined as isotropic. We will now set them for
anisotropic refinement:

Open “Edit atoms”

Select all atoms
Double click any of the selected atoms
Select the radio button “harmonic (isotropic)”
OK;
Set also the number of modulation waves for Cd1 and Cd2 to 1
OK; Yes to quit “Edit atoms”
Double click Refine to refine anisotropic displacement parameters.

The refinement converges to the following R-values:

264 Jana2006 Cookbook, version October 2017

12. Setting up further harmonic waves

Main reflections and first-order satellites now have nice R-factors, but the fit to the higher
satellite orders is not yet good. We need to add higher harmonic positional waves to
properly describe these parameters. There is not universal recipe on the number of harmonic
waves to be set except that their number should almost never be larger than the number of
satellite orders. At the same time, the refined waves should not be too wavy and
unreasonable.

A little bit of testing (try!) shows that an acceptable setting which lowers the R-value and
does not lead to unphysical modulation function is this:

Ni1: 3 waves
Cd1, Cd2, C3, N3, N4: 2 waves
N1, N2, N2b, C2, C2b: 1 wave

Set the number of waves accordingly and run refinement again.

The refinement converges to the following R-values:

The fit is already quite good, but still not perfect. Further testing and checking of the
difference Fourier maps shows that the modulation of ADPs of Cd atoms is required to
improve the fit to the high-order satellites:

265 Jana2006 Cookbook, version October 2017

 Open “Edit atoms”
Select atoms Cd1 and Cd2
Double click the atoms
Set the number of harmonic waves for DP modulation to 1:

Close “Edit atoms and run Refine

The final refinement table looks like this:

13. Plotting the structure

To understand the structure, it can be plotted in a supercell, or various geometric

characteristics can be drawn.

266 Jana2006 Cookbook, version October 2017

Example 6.1: KSm
Modulated structure solved from powder data. Basic steps with powder
refinement, mixed sites in 4d.
Revised: 24 November 2016
K Sm Mo2 O8
Powder data measured with synchrotron radiation
Input files: KSm.dat (powder profile data)
KSm.txt (additional information)
KSm.pdf (article about this compound)
KSm-solution.m40
KSm-solution.m50

1. Creating new jobname

Start Jana2006
“File → Structure → New” opens a file manager
Left pane: locate directory with input files
Right pane: double-click KSm

2. Import Wizard
Select “Various CW format”, NEXT
CW stands for constant wavelength
Type “File name” KSm.dat
Select “Free format of 2th, I, [sig(I)]” and “Another/unknown method”; NEXT
Fill cell parameters 5.5304 5.3019 11.7893 90 90 91.138
The string can be copied from KSm.txt by CtrlC + CtrlV
Change “Target dimension” to 4 and fill q vector 0.56883 -0.12885 0
The string can be copied from KSm.txt by CtrlC + CtrlV
For wavelength type 0.7114 and select “Linearly polarized beam”; NEXT; FINISH
Accept the data in Data repository
Select “Yes, I would like to continue with the wizard”; OK

3. Checking data
[On the screen: Refinement of the powder profile by the le Bail algorithm]
“Show powder profile”
Profile viewer shows experimental powder profile
The powder profile is stored in KSm.m95 and in KSm.m90 in analogy to single crystal data
Leave profile viewer
“Edit refinement commands”; 0 refinement cycles; OK
“Run Refine”
“Show powder profile”
Now the profile viewer shows experimental and calculated powder profile. The calculated
profile is based on default profile parameters. Bragg positions of satellites are green. By
default satellites up to the order 1 are used.
The calculated profile is stored in KSm.prf

267 Jana2006 Cookbook, version October 2017

 Leave profile viewer
“Edit profile parameters”
This opens “Powder options” where profile parameters will be edited
The profile parameters are stored in file KSm.m41

4. Refinement of profile parameters

[On the screen: Powder options]
In the initial state refinement of cell parameters and q-vector is disabled; in page “Profile”
only GW has non-zero value and it is also fixed; in “Corrections” page zero background is
predefined.
In “Profile” page activate refinement of “GW”
In “Corrections” page activate refinement of “shift”
Refinement of background terms is enabled by default
Leave Powder options, save changes
“Edit refinement commands”
Set 100 refinement cycles; OK; “Run Refine”
Refinement converges with Rp~3.7%
The convergence may be slow. You can interrupt the refinement when Rp reaches 3.7% by
button “End” pressed at the moment when LeBail decomposition is not running
“Edit profile parameters”: activate refinement of cell parameters (a, b, c, gamma) and
components (q1,q2) of the modulation vector;
set number of Legendre terms for background to 16
“Run Refine”
Refinement converges with Rp~3.1%
In “Edit profile parameters” switch to the page “Profile” and activate refinement of GU, GV
run Refine
Refinement converges with Rp~2.5%. Waiting for the full convergence is not necessary.
Although the profile viewer shows that profile description could be further improved, for
structure solution this starting point is sufficient.
FINISH; NEXT to open a wizard for the space group test

5. Space group determination

[On the screen: Tolerances for crystal system recognition]
Use default; NEXT
[On the screen: Select Laue symmetry]
Select the highest one – monoclinic; NEXT
[On the screen: Select cell centering]
Select I centering; NEXT
[On the screen: Information about progress of space group tests. It takes some time…]
[On the screen: Select space group];
Select C2/c; NEXT
[On the screen: Final step of the space group test];
Accept the space group transformed into the original cell; Finish
[On the screen: “Do you want to refine profile parameters one more”];
Yes
Now the used symmetry speeds up the profile refinement. Rp is slightly higher because we
don’t use forbidden Bragg positions

268 Jana2006 Cookbook, version October 2017

6. Structure solution
Select “Use Superflip”
Type formula K Sm Mo2 O8; Formula units 2

As in the space group I2/b the minimal site multiplicity is 4 it is clear the Sm and K are either
alternatively occupying the same position or two positions with half occupancy. The first
possibility would change the average density at Sm/K position to a level not to different
from that at Mo position. The average atomic number of Sm/K 40.5 comparing with 42 for
Mo atoms. This means that then we could have problem to distinguish Mo and Sm atom.

Select “Repeat Superflip: Number of runs”, type 10 runs and 3000 cycles
Select “Starting model” = “Patterson superposition map”
Peak from Jana2006 but first run Fourier
“Run solution”

“Accept last solution ” closes the structure solution wizard

“Quit” closes the structure solution wizard and ignores any solution

At this point the powder wizard finishes

The components of the powder wizard can be started separately like
“Parameters → Powder”
“Tools → Powder → Profile viewer”
“Run → Solution”

Start “Edit atoms” or “Edit/View → Editing of M40” and make sure the atomic positions
returned by Superflip are as follows (or symmetry transformed):

Sm1 0.500000 0.250000 0.147357

Mo1 0.000000 0.750000 0.120140

The position of heavy atoms makes situation even more complicated and it induces
additional translation symmetry (1/2,1/2,0) amd (0,0,1/2)

Start “File → Structure → Copy in”

Import M40 and M50 from the structure “KSm-solution”.

This will unify the origin selection with the selection used in the manual.

7. Completing average structure

Atoms in the initial structure model have already modulation functions, which were found in
the electron density map from Superflip. However, for this example it is better to use the
classical approach: calculating first the average structure and then refining modulation
waves from small starting displacements.
Start “Parameters → Powder”, switch to page “Cell”, and check “Use only satellites
corresponding to existing modulation waves” and disable refinement of q vector
components.
Start “Edit atoms” and delete all light atoms (if present)
In our case we will delete all oxygen atoms. Positions of light atoms found from Superflip
269 Jana2006 Cookbook, version October 2017
electron density map may be unreliable.
In “Edit atoms” dialogue select heavy atoms (Sm1 and Mo1), go to “Action → Edit/define”
and delete position modulation wave (if present).
In Refinement options go to page “Basic” and set Number of cycles to 100 and Damping
factor to 0.25
Uncheck “Make only profile matching”
Run refinement
Refinement converges with Rp=22%, R=11%
Activate refinement of occupancy (parameter “ai”) for Sm1 and run Refine
Rp slightly improves, Occupancy of Sm1 changes from 0.5 to 0.4 and ADP of Mo1 becomes
positive. The decrease in occupancy of Sm1 indicates that it could be mixed with potassium.
Open options for Fourier and select difference Fourier map, automatic Scope and default
map orientation. In the “Peaks” page – “Interpretation of displacement waves” change “No.
of harmonics” to 0.
Run difference Fourier
NO (skip procedure for including new atoms)
The fact that both heavy atoms are located at the special positions and with interatomic
vector about (0.5,0.5,0) leads to a false symmetry in the map. However, the true maxima
should form a tetrahedron around Mo1. In the next steps, we will plot Mo1 and neighboring
Fourier maxima in Diamond.
Run “Plot structure”; Select “Draw average structure”;
Press “Add Fourier peaks”; select all maxima; OK; “Draw+continue” to start Diamond
[On the screen: Diamond]
Diamond assigns to the Fourier maxima the lightest chemical element in the formula. For
our formula, this is the oxygen. Sometimes it may be practical to append another light atom
to the end of the formula, for instance hydrogen.
Delete the automatically created structure
With button from the bottom toolbar select Mo1 as the only atom to be plotted
With “Build → Connectivity” focus Mo-O distance and set the limits above 1.5 Å and below
2 Å; OK
Focus Mo1 and plot its coordination with button
Rotate the plot until you recognize a tetrahedron formed by four maxima

270 Jana2006 Cookbook, version October 2017

 In our example we can make a tetrahedron from Max2 and Max3 as well as from Max1 and
Max6 and thay have different bond lengths. The second tetrahedron is a product of the false
symmetry induced by additional translation symmetry mentioned above In our case it is
difficult to say which one is better so we will exclude the second one because Max6 may be
too weak maximum (maxima are numbered sequentially by their intensity).
Start “Parameters → Atoms → New” and include the maxima previously selected in
Diamond (Max2 and Max3 in our case) as O1 and O2.
Run refinement
Refinement converges with Rp=13%, R=5%
Start “File → Structure → Save As” and make a backup copy of the average structure, for
instance KSm_ave

8. Refinement of modulated structure

Start “Edit atoms” and set one position modulation wave for Sm1
Run refinement
Add one position modulation wave for Mo1
Run refinement
Refinement converges with Rp=12%, R(main)=5% and R(sat)=15%.
Run Contour, press “New plot”, OK
Select “Calculate new map”
In the “Basic” page select observed Fourier map
In the page “Scope” choose “by a central point”, clear the checkbox “Use default map
orientation”, select first axis x1 and second axis x4 (i.e. the x1-x4 section), type Sm1 for the
Center and “2 1 1” for the section scope.

271 Jana2006 Cookbook, version October 2017

Press “Sum on” and check “x2” and “x3”
Press “Atom edit” and add Sm1 to the list of atoms which will be indicated in the map
The resulting modulation function of Sm1 looks continuously. However, when we follow the
line by mouse cursor the Density window in the lower right corner displays fluctuation of
density between 68 and 51 e-A-3. Because position of potassium does not appear in the
average structure and the occupation of samarium is lower than 0.5 we assume that
samarium mixes with potassium. In modulated structure, classical disorder is often replaced
by distribution of elements along x4. We can speculate that the area with lower electron
density corresponds to the interval (in x4) where potassium exists instead of samarium.
However, better background for this consideration will be difference Fourier map.

[On the screen: Contour window]

Press “New plot”, select “Calculate new ones”, OK

272 Jana2006 Cookbook, version October 2017

In Fourier commands select difference Fourier map, OK
In Contour window that has automatically appeared press “Sum ON” and make summation
along “x3” and “x2”

The positive maximum in the difference map indicates region with lack of electron density.
This is the area where fully occupied samarium should be present (instead of the partially
occupied one). From this map we can conclude that the definition interval of potassium has
centre in x4=0 and width 0.5. Samarium would occupy a complementary interval with centre
at x4=0.5 and width 0.5.
Leave Contour and start “Edit atoms”
Right-click Sm1 and select “Split atomic position”
In the dialogue that follows:
Type “K1” for “Name of the split atom”
Select “K” for “Atomic type”
Type “0.5” for “Relative occupancy of the split atom”
Type “0 0 0” for “Shift of the split atom”
Clear “Generate restrict command for Refine”
Apply

273 Jana2006 Cookbook, version October 2017

By this way we have duplicated Sm1, assigned the name K1 and chemical type K to the
duplicated atom and reduced occupancy of both atoms by 1/2.

[On the screen: list of atoms in “Edit atoms”]

Select Sm1 and K1, go to “Action → Edit/define atoms” and check “crenel”

Switch to the page “Edit”, ensure that “ai” = 0.5 and its refinement is disabled
Press “Occupancy”
Type 0.5 for “delta” and 0.5 for x40, OK, OK
By this way we have defined two identical crenel functions for both atoms Sm1 and K1. Note
that the occupancy “ai” was recalculated back from 0.25 to 0.5
[On the screen: list of atoms in “Edit atoms”]
Unselect Sm1
Right-click K1, go to “Action → Edit/define atoms”
Switch to page “Edit”; press “Occupancy”, change x40 from 0.5 to 0; OK; OK
Close “Edit atoms” and save changes

Start Contour and plot the previous difference Fourier map

Make summation
Press “Atoms edit and add K1 to the list”
Make the same for observed Fourier map
The section shows complementary modulation functions of Sm1 and K1

274 Jana2006 Cookbook, version October 2017

Open Refinement options, go to the page “Various”, press Restrictions
In the textbox type “Sm1 K1”
Check “Coordinates”, “Modulations” and “ADP parameters”
For “Occupancies” select “not restricted”; press “Add”
This restriction makes the two atoms Sm1 and K1 identical except the definition interval in
x4
Leave Refinement options, save them without starting refinement (YES)
Create backup copy “KSm_crenel”
Run Refine
The refinement converges with Rp=6%, R(main)=2.2% and R(sat)=6.5%
Add one position modulation wave for O1 and O2
Start “Parameters → Powder”, go to the page “Cell” and enable refinement of “q1” and
“q2”
The components of q vector were fixed because they can be refined only when modulation
waves are present due to the option “Use only satellites corresponding to existing
modulation waves”
Run Refine
The refinement converges with Rp=6%, R(main)=2% and R(sat)=5.5%
Plot observed and difference Fourier map for Sm1

275 Jana2006 Cookbook, version October 2017

 x1-x4 difference x1-x4 observed x2-x4 observed

Contour step in the difference map is 0.1. The residua in the difference Fourier map are smaller
and in different places than before. The position modulation now occurs in x2-x4 section. The
previously observed modulation in the x1-x4 section was caused by pseudosymmetry.

9. Better profile description

Create backup copy KSm_mod
Start “Tools → Powder → Profile viewer” and examine the first two peaks of the profile

The profile width can be probably better described with anisotropic strain broadening
Start “Parameters → Powder”, go to the page “Profile”
Select “Tensor method”
Press “Edit tensor parameters”
Press “Refine all”
Leave “Powder options”
In Refinement options go to page “Basic” and select “Use Marquart technique” with the
default “Fudge factor” 0.001
Run Refine
and fix it to zero in “Edit tensor parameters”.
The refinement converges with Rp=3.8%, R(main)=1.7% and R(sat)=2.9%

276 Jana2006 Cookbook, version October 2017

Check Powder profile

The width of reflections is now described correctly

Open “Powder options”, go to the page “Profile”
Select “Pseudo-Voight” and activate refinement of “Lx” and “Ly”
Run Refine
The refinement converges with Rp=3.0%, R(main)=1.4% and R(sat)=2.1%
Some St tensor component may cause oscillations. A careful analysis would be needed to
select the St components really necessary for profile description. We can for instance fix the
components with low parameter-to-sigma ratio (using the button “Show p/sig(p)” to see the
ratio)
Plot difference Fourier for Sm1, section x1-x4

The section now displays very small residual density. Depending on the fixed St components
the resulting section may look differently.

Start “Plot structure” and try to plot the structure in Diamond

Sufficiently large area must be selected for the plot because in Diamond we can only plot the
supplied atoms. No symmetry transformation (rotation as well as translation) is allowed in
Diamond because it does not use modulation functions. The recommended way is to add
atoms supplied by Jana2006 using button , connect them using button and then
make only selections and rotations.

277 Jana2006 Cookbook, version October 2017

Alternatives
The fact that the Sm/K position is half occupied by samarium and potassium makes its
scattering power almost identical with Mo atoms. This induces a pseudo-translation symmetry
(1/2,1/2,0) (0,0,1/2) which strongly affects the solution. The chosen approach is not the only
possible one. Another way would be to start – after completing the average structure – with
refinement of position modulation of both Mo1 and Sm1 together with harmonic occupation
wave of samarium. With this way the “false” modulation in the x1-x4 section would not appear.

278 Jana2006 Cookbook, version October 2017

Example 7.1: Ephedrine
Commensurately modulated structure
Revised: 1 October 2017

C10H16NOCl

Single crystal data measured with Bruker four-circle diffractometer

Input files: Ephedrine.hkl, Ephedrine.ins
Frame scaling, absorption correction: done

1. Creating new jobname

Start Jana2006
“File → Structure → New” opens a file manager
Left pane: locate directory with input files
Right pane: double-click Ephedrine

2. Import Wizard
Select “reflection file corrected for LP and absorption”; NEXT
Select “SHELX on I”; File “Ephedrine.hkl” NEXT
The measurement was done in a supercell using three indices
[On the screen: complete/correct experimental parameters]
Fill supercell parameters as used in data collection. They can be copied (Ctrl-C) from the file
Ephedrine.ins and pasted (Ctrl-V) to the “Cell parameters” textbox; NEXT
[On the screen: Define the reference cell/split by twinning]
Change target dimension for 4
Define modulation vector 0 0 1/4
Press “Define transformation matrix” and fill the matrix array by

100
010
0 0 1/4

NEXT; FINISH

3. Data Repository
The line in data repository window corresponds to one data set.
OK; YES to accept the data set

279 Jana2006 Cookbook, version October 2017

4. Symmetry Wizard
[On the screen: “Tolerances for crystal system recognition”.]
Leave all settings default; NEXT
[On the screen: Select Laue symmetry]
Select Orthorhombic Laue symmetry; NEXT
[On the screen: Select cell centering]
Select P centered unit cell; NEXT
[On the screen: Select space group]
Select P212121; NEXT
[On the screen: Select superspace group]
P212121(00g)000
[On the screen: Final step of the superspace test]
Accept the superspace group transformed into the original cell: P212121(00g)000; FINISH
[On the screen: Define superspace group]
OK

5. Creating refinement reflection file

In this step we have M95 (i.e. the data from diffractometer converted to Jana format) and
symmetry. For refinement we need M90 containing data set merged by symmetry and with
discarded extinct reflections.
NEXT to confirm threshold 3sigma
[On the screen: settings for merging of reflections]
Leave all settings default; NEXT
Accept the data block; FINISH

6. Structure Solution Wizard

[On the screen: window of Structure solution]
In “Formula” textbox type list of chemical elements for C10H16NOCl: C H N O Cl
Select “Superflip”, “Peaks from Jana2006 but first run Fourier”; Run Solution
[On the screen: window of Superflip replaced later on by listing of Superflip]
Superflip converges (after noise suppression) with R value between 20 and 30%.
Note that we don’t fix random seed and the exact result of Superflip will differ between runs
Close the listing of Superflip
Press “Accept last solution”

7. Verification of the structure solution

Start EditM50; go to page “Composition”; press “Formula from M40”
[On the screen: Information about formula]
It should display formula C10 H0 N O Cl which means that all atoms except hydrogen atoms
were localized. In the case that some atoms are missing they can be found from subsequent
refinement and difference Fourier maps. In the case that the Superflip gives some additional
false atoms they usually are detected in the refinement as atoms having non-realistic (too
high) isotropic ADP and they will be automatically eliminated by putting their occupancy to
zero. The atoms being eliminated can be later deleted in “Edit atoms” with help of button
“Select rejected”.
Start “Plot structure” to check assignment of chemical types (by comparing with the
structure formula given at the beginning of this chapter). Use the option “Draw average

280 Jana2006 Cookbook, version October 2017

 structure”.
If some atom in the plot looks like a false maximum it should be deleted by “Atoms edit”

8. Refinement of the modulated structure

Start Refine with default parameters.
You should get R values about 9.3%, 7.5%, 7.6% and 31.2%.
If some atoms are missing they should be found in difference Fourier, added to the structure
model and the refinement should be repeated
Start “Edit atoms”, select all atoms, change ADP to harmonic ones and repeat the
refinement.
You should get R values about 8.4%, 6.4%, 6.5% and 31.4%.

9. Adding hydrogen atoms

Start “Edit atoms”, select all carbon and nitrogen atoms and add hydrogen atoms
automatically.
For nitrogen atoms automatic procedure is not reliable, hydrogen atoms will be added
manually
[On the screen: Adding of hydrogen atoms for N1]
Select “Tetrahedric”, change number of neighbors to 2
You can verify the neighbors by focusing the textbox with the first neighbor and clicking
“Select neighbors”
Define names of hydrogens, for instance H1N1 and H2N1
“Apply”

Run EditM50 and go to page Composition, press Formula from M40. The found formula
should be C10 H15 N O Cl which means that one hydrogen atom bounded to O is missing.
This atom will be localized later from the Fourier map.
Run refinement
Refinement converges with R value about 6.2%, 3.4%, 4.9% and 31.6%.

10. Localization of the remaining hydrogen atom

Right-click the icon of Fourier
In page “Basic” select “Map type” as “F(obs)-F(calc) – difference Fourier”
In page “Basic” select “Apply sin(th)/lambda limits” and set tli limits from 0 to 0.5
This will suppress noise originated from high order reflections, to which H atoms contribute
negligibly
In page “Scope” select “Automatically” and use the default parameters for “Independent
parallelepiped”
Run Fourier.
[On the screen: Do you want to start the procedure for including of new atoms?]
Answer yes.
One of the dominating maxima should have a reasonable hydrogen distance to the oxygen
atom O1. Add it as a new hydrogen atom (named e.g. H1O1).
Right-click the icon of Refine; select page “Various”; press the button “Keep commands”
Restrict ADP parameter of the newly introduced hydrogen atom.

281 Jana2006 Cookbook, version October 2017

 Add the new keep command and start the refinement procedure.
Refinement converges with R value about 6.0%, 3.2%, 4.9% and 31.60%.
Verify by plotting that the hydrogen in OH has a reasonable position
Save the structure with a new name e.g. “Ephedrine-incomm” but continue with the old
one. This structure will be used for a final comparison with two commensurate cases.

11. Introduction of the commensurate option and twinning

Start EditM50, go to page Cell.
Select “Commensurate case” and “Use simple supercell”
In text box “Supercell” type “1 1 4”
Press button: “Select supercell group” and select Tzero = 0 + n/4 which represent
monoclinic symmetry with two-fold screw axis along a.
As the supercell symmetry is lower than orthorhombic we may have two twinning domains
related either by two fold axis along b or along c direction.
Select twinning and define the twinning matrix as a 180 degrees rotation along c axis.
The selection of the commensurate model leads to different Laue symmetry and different
extinctions and therefore we should generally redo the creation of the reflection file for the
refinement. But for better comparison of tested models we shall skip this point.
YES to the question “Do you want to rewrite M50 file ?”

“DO you want to re-create refinement reflection file now?” NO in order to skip re-creation
of the reflection file
Go to “Parameters/Twin fractions” and allow refinement of twvol2
Run Refine.
Refinement converges with R value about 3.8%, 2.9%, 4.1 and 8.0%.
Save the structure with a new name e.g. “Ephedrine-comm-1” but continue with the old
one.
This structure will be used for a final comparison with two commensurate cases.

12. Commensurate case for t0=1/16

Repeat the point 11 for Tzero = 1/16+n/4.
NO in order to skip re-creation of the reflection file

282 Jana2006 Cookbook, version October 2017

 Run Refine.
Refinement converges with R value about 3.8%, 2.9%, 4.1 and 8.0%.
The result is similar like in the previous commensurate model
Save the structure with a new name e.g. “Ephedrine-comm-2”

13. Incommensurate model with overlap option

Go to “File → Structure → open” and select files containing the incommensurate
refinement.
In the incommensurate refinement we are not taking into account the fact that main
reflections are overlapped with satellites of fourth order and the satellites of the first order
are overlapped with those of the third order. Even more serious are overlaps of the second
order satellites as they have in average similar intensities.
For the commensurate structure the program calculates structure factors as a summation
over really existing t sections, taking into account t0 value and with symmetry elements
being closed in the supercell. Thus we get the same result for reflections:

(h,k,l,m), (h,k,l+1,m-4), (h,k,l-1,m+4) e.t.c.

On the other hand, in incommensurate structure model overlaps must be determined based
on the proximity of reflection angles

Start Refinement options

Select page “Modulation” and activate overlap option “Closest reflections”; OK
Run Refinement

Refinement converges with R value about 3.8%, 2.9%, 4.1 and 8.0%.

Conclusion: Modulations are not very strong and therefore all three models give very similar
results

Further possible steps:

- for selected model second modulation wave can be added, but it has almost no influence on
results

- we can deactivate twinning and verify by refinement if it was really needed (it was!)

- hydrogen of OH group probably should not have individual modulation waves. They can be
made the same like modulations of the parent atom in “Refinement options → page Various →
Restrictions”

- In the incommensurate refinement the overlaps detected by the program can be printed in the
refinement listing if we activate the option in “Refinement options → page Select/Listing →
Print twin/overlap details”

283 Jana2006 Cookbook, version October 2017

284 Jana2006 Cookbook, version October 2017
Example 7.2: CrPOcom
Solution of the low temperature commensurate phase of Cr2P2O7.
Revised: 20 April 2017
Cr2P2O7

Single crystal data measured with Oxford Diffraction four-cycle diffractometer

Input files: CrPOcom.hkl, CrPOcom_red.sum
Frame scaling, absorption correction: done with software of the diffractometer

1. Creating new jobname

Start Jana2006
“File → Structure → New” opens a file manager
Left pane: locate directory with input files
Right pane: double-click CrPOcom

2. Import Wizard
Select “Known diffractometer formats”; NEXT
Select “Oxford Diffraction – CCD”; NEXT
Define basic input file: select Input from “sum” file; OK
Leave all settings unchanged; NEXT
Leave all settings unchanged; NEXT
For absorption correction select “None or done before importing”; NEXT
FINISH

3. Data Repository
The line in data repository window corresponds to one data set.
OK; YES to accept the data set

4. Symmetry Wizard
Use default tolerances for crystal system recognition
Select monoclinic Laue symmetry
Select C centering
Select space group C2/m
Select superspace group C2/m(a0g)0s
Accept the superspace group in the original cell

5. Creating refinement reflection file

Follow default steps of the wizard

6. Structure solution wizard

Type the formula Cr2 P2 O7, two formula units
Select “Superflip” and “Peaks from Jana2006”; “Run solution”
Superflip converges with R around 25%, confirms the selected superspace group

285 Jana2006 Cookbook, version October 2017

 Close the listing of Superflip
“Accept last solution”
“Quit” to leave the structure solution wizard

7. Workaround to get the same atomic positions like in this cookbook

Superflip returns arbitrary symmetry equivalent origin position. This would complicate
reading of the cookbook because atomic positions and Fourier sections could be different
(shifted, transformed) than it is described in the text. At this point, we shall use a saved
Superflip solution compatible with the cookbook:
With your file manager:
copy superflip.m40 to jobname.m40 (“jobname” is name of your current job)
copy superflip.m80 to jobname.m80
copy superflip.m81 to jobname.m81
Jana2006 can be open during copying as it always unlocks the files when going to the basic
window
M40 contains structure model from Superflip, M80 contains structure factors found by
Superflip, M81 is the electron density map calculated from M80

Plot the average structure with Diamond

From the plot, it is clear that the two positions of the bridging oxygen O2 are probably
separated in the modulated structure with a crenel function

8. Inspecting modulation functions returned by Superflip

Start Contour and “New plot”
Select “Show de Wolff’s sections” and set following parameters:

286 Jana2006 Cookbook, version October 2017

Press “Define atoms for de Wolff’s sections”
Add all atoms, for each of them select sections x,y,x.
OK; OK
In Contour, plot the section using “W+”, “W-“ and “From list”

Oxygen O2 has x2=-0.07, its symmetry equivalent position has x2=+0.07. Therefore we can
define a crenel function of the width 0.5 and the center 0.75 which will be transformed by
the symmetry to the crenel function with the same width, the center 0.25 and x2=+0.07. This
will describe “skipping” of the bridging oxygen between two positions separated by the shift
in x4.

Similarly like in example 5.3.2 (incommensurate phase of Cr2P2O7), we will combine crenel
functions with position modulations expressed by Legendre polynomials.

287 Jana2006 Cookbook, version October 2017

Leave Contour, start “Edit atoms”
Edit O2: in “Define” page check “crenel” (it will add one occupation wave)
Set delta=0.5 and x40=0.75; OK
Select “legendre polynomials in crenel interval”
Leave Edit atoms (save the results)
Plot the same section with Contour, using the button “From list”
In “Atoms edit” (of Contour) add atoms O2#s4

The Contour plot now shows the two crenel functions related by symmetry operation
x1 -x2 x3 x4+1/2

288 Jana2006 Cookbook, version October 2017

Using the button “From list” in Contour, plot similar x2-x4 sections for P1

P1 needs more modulation waves (will be added later)

Using the button “From list” in Contour, plot x1-x4 sections for Cr1

Cr1 should be described with by crenel function with center 0.25 and width 1 combined with
positional modulation to create discontinuity at x4=0.75.

Similarly like for O2 define crenel function for Cr1: x40 = 0.25, delta = 1
Select “legendre polynomials in crenel interval”

289 Jana2006 Cookbook, version October 2017

Plot the same section with Contour:

Using the button “New plot” in Contour plot x2-x4 sections for O1

O1 has no discontinuity and only small positional modulation

290 Jana2006 Cookbook, version October 2017

 Using the button “New plot” in Contour plot x1-x4 sections for O3

O3 should be described with sawtooth function.

Similarly like for O2, define crenel function for O3: x40 = 0.69, delta = 1
Select “legendre polynomials in crenel interval”
Plot the same section in Contour:

9. Refinement
With commensurate structures, crenel function can be used but its center and width cannot
be refined, as their changes give zero derivations, which would lead to singularity problems
during refinement.
Start “Editm50”: check “Commensurate case”, “Simple supercell”, supercell 3 1 2, Tzero = 0
Follow wizard for creating the refinement reflection file
Start Refinement commands: 100 cycles, damping 0.5
291 Jana2006 Cookbook, version October 2017
Refine
Refinement converges with 18%, 8%, 13%, 46% and 56% for all, main, 1st order, 2nd order
and 3rd order (31 refined parameters).
Add the second position modulation wave for all atoms except O2
Refine
Refinement converges with 8%, 4%, 6%, 15% and 48% for all, main, 1st order, 2nd order and
3rd order (48 refined parameters).
Define harmonic ADP for all atoms
Refine
Contour: plot relevant sections for O2, O3 and Cr1

O2 O3 Cr1

Add second position modulation wave for O2

Refine
At the message about singularity, choose “Skip warnings”
Open refinement listing and use Use “Go to → Singularity report”
For commensurate refinement, the second modulation wave for O2 cannot be used. It would
introduce more refinement parameters than the structure would have in a supercell
description.
Delete the 2nd position modulation wave for O2
Add the 3rd position modulation wave for all atoms except O2
Refine (with skipped warnings)
Adding the 3rd position modulation wave improved significantly the fit!
Look at the singularity report in the refinement listing
The list of blocked parameters contains either “cos3” terms (for harmonic modulation) or
“ort6” terms (for Legendre modulation). This means that the second halves of the third
modulation waves are forbidden (for the same reason given for O2).
“Refinement commands → Various → Fixed commands”
Use “Set individual” and set “*cos3[*]” to zero
Use “Set individual” and set “*ort6[*]” to zero
“?” and “*” are wildcards.
Refine

292 Jana2006 Cookbook, version October 2017

 Refinement converges without warning messages to 2.7%, 2.5%, 2.6%, 2.8% and 4.1% for
all, main, 1st order, 2nd order and 3rd order (72 refined parameters).

Add one harmonic ADP modulation wave for all atoms

Refine
Refinement converges without warning messages to 2.1%, 2.2%, 1.9%, 2.1% and 2.6% for
all, main, 1st order, 2nd order and 3rd order (104 parameters).

“Parameters → Extinction”: set isotropic Lorentzian extinction

Refine
Refinement converges without warning messages to 2.1%, 2.1%, 1.9%, 2.1% and 2.5% for
all, main, 1st order, 2nd order and 3rd order.
Number of refined parameters: 105

10. Binding t40 of O2 and O3

Start Grapht.
With New/Edit: “Parameter to draw” position – y; “p” from -0.3 to 0.1; t from 0 to 0.2;
Name P1; step in t 0.0001
Press “New Item” and define the same for O2
Press “New Item” and define the same for O3

OK to plot

293 Jana2006 Cookbook, version October 2017

Move mouse cursor over the lines to see which atoms they are belonging to
The discontinuity in O3 occurs at slightly different t then the discontinuity in O2. However,
they belong to the same P1 tetrahedron and therefore they must skip in phase.
Close Grapht

Options for Dist: in “Basic” page set listing “With symmetry code”; in “Select atoms” page
set central atom P1, neighbor atoms O2, O3.
Start Dist
Open the listing of Dist
Atom P1 is coordinated with O2 and O3, both in the basic symmetry position x,y,z
The upper limit of the crenel interval should be the same when expressed in t:
tmax(O3) = tmax(O2)
t40(O3) + delta(O3)/2 = t40(O2) + delta(O2)/2
t40(O3) + 0.5 = t40(O2) + 0.25
t40(O3) = t40(O2) - 0.25

Refinement commands: in page “Various” select “Equations” and define

t40[O3]=t40[O2]-0.25
Refine
The R values are almost unchanged
An equivalent solution could be found with lower limits if the crenel interval (t40(O3) =
t40(O2) + 0.25)
Open refinement listing
Look at “Equations defined by user” where the user equation is expanded

Start Grapht and verify that the discontinuity in O2 nad O3 now occurs at the same t

294 Jana2006 Cookbook, version October 2017

11. Transformation to supercell
Start “Tools → Transformations → Go to supercell structure”
Follow the wizard for creating refinement reflection file
Continue with the new (supercell) structure
EditM50: check cell parameters (3x1x2 supercell) and symmetry (I2/c)
Refinement commands: set 0 cycles
Refine
Refinement (with zero cycles) confirms overall R value found previously in the
commensurate model
Number of structure parameters is 151. This means we were able to get satisfactory results
in commensurate refinement saving 46 parameters.
Supercell model can be used for plotting of the structure and for calculation of geometry

“File → Structure → History”: return to the commensurate model

“File → Save As”: make a backup copy
Set three ADP modulation waves to all atoms
Refine – skip warnings
The number of refined parameters is now much larger but no further improvement o R
values occurs. By adding modulation waves and fixing forbidden components to zero we
should finally reach the same number of structural parameters in the commensurate and
supercell model.

295 Jana2006 Cookbook, version October 2017

296 Jana2006 Cookbook, version October 2017
Example 7.3: PhenanTin
Commensurate and supercell description of a five-fold superstructure.
Revised: 20 April 2017

Bis(aquachlorotriphenyltin 1,10-phenantroline), C150 H125 N10 O5 Cl5 Sn5, Z=4

Single crystal data measured with four-cycle diffractometer SuperNova

Input files:
PhenanTin.hkl, PhenanTin.cif_od (diffraction data indexed in the supercell - see later)
phenantroline-no-symmetry.m45 (pattern molecule phenantroline)
phenantroline-no-symmetry.jpg
phenyl-no-symmetry.m45 (pattern molecule phenyl)
phenyl-no-symmetry.jpg
supercell_keep.txt (pre-prepared keep commands for supercell refinement)

Chemical diagram:

Introduction
Structure of this compound has been published in Journal of Science and Technology in the
Tropics (2005), 1, 157-163, as a commensurately modulated structure. The data from Nonius
KappaCCD diffractometer contained strong main reflections and weak first and second order
satellites with ratio of intensities 1000 : 0.157 : 0.075. Later on the compound was remeasured
with much more intensive source of the Rigaku Oxford Diffraction SuperNova. In the new
experiment, the satellites were strong and the structure could be solved in a five-fold supercell
as a classical three-dimensional structure. This example uses data of SuperNova and compares
possibilities of the commensurate and supercell structure refinement.

297 Jana2006 Cookbook, version October 2017

Reciprocal space with peak positions projected along b*. Black: unit cell for commensurate
refinement; Red: unit cell for supercell refinement.
Relationship between unit cells in direct space:
1 0 1 2/5 0 1/5
Tcomm->sup = �0 −1 0�; Tsup->comm = � 0 −1 0�
3 0 −2 3/5 0 −1/5

1. Creating new jobname

Start Jana2006
“File → Structure → New” opens a file manager
Left pane: locate directory with input files
Right pane: double-click PhenanTin

2. Import Wizard
Select “Known diffractometer formats”; NEXT
Select “Oxford Diffraction – CCD”; NEXT
Define basic input file: select Input from “cif_od” file; OK
Leave all settings unchanged; NEXT
Leave all settings unchanged; NEXT
For absorption correction select “None or done before importing”; NEXT
FINISH

3. Data Repository
The line in data repository window corresponds to one data set.
OK; YES to accept the data set

4. Symmetry Wizard
Use default tolerances for crystal system recognition
Select monoclinic Laue symmetry
Select P centering

298 Jana2006 Cookbook, version October 2017

 Select space group P21/n
Although it is not first in the list, FOM is almost the same like for P2/n and the button
"Details" shows that the discarded reflections are very weak
Accept the space group transformed into the original cell; FINISH

5. Creating refinement reflection file

Follow default steps of the wizard
Rint ~ 4.8% for 24879 reflections

6. Structure solution wizard

Type the formula C150 H125 N10 O5 Cl5 Sn5, four formula units
Density should be 1.53 g.cm-3
Select “Superflip” and “Peaks from Jana2006”; “Run solution”
Superflip converges with R around 20% and confirms the selected space group
“Accept last solution”
Plot the structure

The structure model returned by charge flipping contains most of the expected structure
and it could be finished by adding maxima from difference Fourier and possibly by applying
geometry constraints imposed during the refinement. However, we would like to try
commensurate refinement in order to see its pros and cons.

7. Data import for commensurate refinement

As the structure handled in the 3d supercell cannot be simply transformed to the
superspace, we have to re-create the structure. We will use the same job name.

“File → Structure → New” opens a file manager

Left pane: locate directory with input files
Right pane: double-click PhenanTin
YES to overwrite the existing job name
Select “Known diffractometer formats”; NEXT
Select “Oxford Diffraction – CCD”; NEXT
Define basic input file: select Input from “cif_od” file; OK
[On the screen: Complete/correct experimental parameters]

299 Jana2006 Cookbook, version October 2017

 Leave all settings unchanged; NEXT
[On the screen: Define the reference cell/split by twinning]
Change target dimension to 4
For "1st modulation vector" type 0.2 0 -0.2
The modulation vector has several possible settings. For instance, the original publication
used (0.2,0,0.8). However, the short q-vector as defined here is the most logical selection.
Press "Define transformation matrix"
[On the screen: Define the transformation matrix]
Define the following 3x3 matrix: 2/5 0 1/5; 0 -1 0; 3/5 0 -1/5
“inp” means the input indices; “float” means the transformed indices. They can generally be
real numbers, in such case the program tries to index them as satellites
OK
The transformed experimental unit cell parameters should be as follows:

NEXT
The wizard transforms indices according to the defined transformation matrix. Resulting
integer indices are sorted as main reflections while the others as satellites defined by the q-
vector.
For absorption correction select “None or done before importing”; NEXT
FINISH
OK; YES to accept the data set in the Data repository

8. Symmetry Wizard
Use default tolerances for crystal system recognition
Select monoclinic Laue symmetry
Select P centering
Select space group P21/c
Although it is not first in the list, FOM is almost the same like for P2/n and the button
"Details" shows that the discarded reflections are very weak
NEXT
Select superspace group P21/c(a0g)0s; NEXT
Accept the superspace group transformed into the original cell; NEXT

300 Jana2006 Cookbook, version October 2017

9. Creating refinement reflection file
Follow default steps of the wizard
Rint ~ 4.8% for 24879 reflections

10. Structure solution wizard

Type the formula C30 H25 N2 O Cl Sn, four formula units
Density should be 1.53 g.cm-3
Select “Superflip” and “Peaks from Jana2006 – but first run Fourier”;
For “Starting model” select “Patterson superposition map”
“Run solution”
Superflip converges with R around 22% and confirms the selected superspace group
“Accept last solution” to leave the structure solution wizard
Start "Plot structure"
Select "Draw average structure" and press "Draw+return" or "Draw+continue"
In Diamont start "Build → Get molecules"

The structure model obtained from charge flipping contains fragments of phenyls and
phenantroline, Sn, Cl and water oxygen assigned as carbon C1. In commensurate description
we have only one symmetry independent Sn.

11. Initial commensurate refinement

As the only special position in the superspace group P21/c(a0g)0s is the inversion center,
which is not compatible with the assumed shape of the molecule, all atoms should be
localized in general positions. Thus one atom in the commensurate structure represents five
atoms in the supercell structure, i.e. 15 positional and 30 ADP parameters. Therefore, each
atom in the commensurate structure can have maximally two positional (12 parameters)
and two ADP (24 parameters) modulation waves (each wave of order n represents two
periodic components – sin2πnx4 and cos2πnx4), which make together with the basic position
and the basic ellipsoid 3+12=15 positional and 6+24=30 ADP parameters.
At this stage all atoms from the charge flipping model already have one position modulation
wave interpreted from the electron density map provided by Superflip.

Start "Edit atoms"; change chemical type of C1 (may be also N1) to oxygen and its label to
O1;
OK; YES to save changes
Start "EditM50"
Check "Commensurate case"
Clear "Use simple supercell"
Press "Define supercell matrix" and type 1 0 1 ; 0 -1 0 ; 3 0 -2
301 Jana2006 Cookbook, version October 2017
 This is transformation from commensurate to supercell parameters
OK
Press "Select supercell group"
The program lists all supercell space groups, which can be obtained by changing the tzero
value. As we know from the preliminary calculations that the symmetry in the supercell
should be P21/n, we should select one of the two equivalent possibilities.
Select Tzero 0+n/5; OK
OK; YES to rewrite M50; YES to re-create refinement reflection file
Follow steps for merging the reflections; FINISH
Now we have set up the commensurate refinement. In our case, it means that modulation
functions will only have 5 definition points.
Start refinement
The refinement converges with Rall~38%, Rmain~18% (your numbers can be different)
Start "Edit atoms"
Double-click Sn1 and change number of position modulation waves to 2
OK; OK; YES to save changes
Start Refinement options; set 100 refinement cycles; OK; YES+START
The refinement converges with Rall~24%, Rmain~14%, Rsat1~27% and Rsat2~31%
These numbers can be different because Superflip may have returned different number of
carbon atoms. However, if R values at this point are significantly larger, for instance Rall >
26% or Rsat2 ~ 50%, please do not continue and repeat structure solution (return to the
point 10.).

12. Structure completing from difference Fourier

Start Fourier options; select the page "Basic" and "Map type" difference Fourier
OK; YES+START
YES to start procedure for including new atoms
Activate “Skip peaks being too close to existing atom” and set 1.1 Å
Include maxima with reasonable C-C distances named C* (i.e. the next free carbon label).
FINISH; YES to include new atoms
We don’t need a complete structure, just some atomic positions. For correct placement of
molecules in next steps, we need to recognize phenantroline and phenyl. For recognizing N
atoms of phenantroline, it is useful to plot “ellipsoids” although we refine isotropic ADP:
smaller rings will indicate heavier atoms.

13. Preparation for rigid body description

Refinement of a commensurate structure with very strong satellites may be difficult. In this
example, the refinement could easily finish in a false minimum with distorted geometry of
phenyls and phenantroline. Therefore, for the initial refinement we will use fixed rigid bodies
pre-defined in M45 files.
Atoms used for definition of a rigid body position must belong to the same fragment without
need of symmetry transformation. In the next step, we will manage that atoms in M40 are
properly transformed.
Start "Edit atoms"; select all atoms
“Action” → “make symmetry contiguous motifs”
OK; YES to rewrite changes
Start "Plot structure"
Select "Draw average structure"; press "Draw+continue"

302 Jana2006 Cookbook, version October 2017

 In Diamond use "Build → Add all atoms"
This adds all atoms from the structure model without applying symmetry.
Start "Build → Connect atoms"
The plot may differ for actual solutions but it should display all atoms important for
introduction of rigid body of phenyl and phenantroline. In our case, the plot also includes
some spurious maxima, but it does not matter.
For each phenyl and for the phenantroline note labels of three atoms, which will be used
for placing complete phenyl and phenantroline templates from m45
In case of phenantroline note also orientation of the fragment because atoms (here) C2 and
C3 must fit with proper N1 and N2 of the template

Quit Diamond or let it open for a later reference (depends on style of work)

14. Rigid body templates from M45

Start "File → Structure → Save As"
Make backup copy of the structure
Start "Parameters → Molecule → New molecule"
Select "Model file"
For "Name of molecule" type "Ph"
Press "Browse" and select "phenyl-no-symmetry.m45"
Keep the default scaling distance
Press the button “Show the model molecule”
[On the screen: Windows opens jpeg plot of the template in a default viewer. Note that the
atom labels of the template phenyl end with "p" in order not to mix with any atoms of the
structure. Close the jpeg plot.]
NEXT
[On the screen: "Specify atomic types in the model molecule:" dialogue for matching
chemical types in the template molecule and in the current structure.]
Keep the suggested chemical types; NEXT
[On the screen: Define the molecular reference point]
Select "Geom. center"; NEXT
[On the screen: Define and complete molecular position #1]
Pair the following atoms (your labels will differ!)
C1p with C28
C2p with C12
C6p with C17

303 Jana2006 Cookbook, version October 2017

For "Maximal coincidence distance" type 0.8
Press "Show coinciding atoms"
The expected coincidence ratio in our example is 6/11 (in case that all carbon atoms were
previously localized from the difference Fourier)

Press "Apply+Next position"

[On the screen: Define and complete molecular position #2]
Pair atoms:
C1p with C8
C2p with C10
C6p with C13
The coincidence ratio (in our example, it may differ) will be 5/11
Press "Apply+Next position"
[On the screen: Define and complete molecular position #3]
Pair atoms:
C1p with C11
C2p with C6
C6p with C20
The coincidence ratio (in our example, it may differ) will be 6/11
Press "Apply + End"
Now we continue with phenantroline
Start "Parameters → Molecule → New molecule"
Select "Model file"
For "Name of molecule" type "Trol"
Press "Browse" and select "phenantroline-no-symmetry.m45"
Keep the default scaling distance
Press the button “Show the model molecule”
[On the screen: Windows open jpeg plot of the template in a default viewer. Note that the
atom labels of the template phenantroline end with "t" in order not to mix with any atoms
of the structure. Close the jpeg plot.]
NEXT
[On the screen: "Specify atomic types in the model molecule:" dialogue for matching

304 Jana2006 Cookbook, version October 2017

chemical types in the template molecule and in the current structure.]
Accept suggested chemical types; NEXT
[On the screen: Define the molecular reference point]
Select "Explicit"; type N1t; NEXT
[On the screen: Define and complete molecular position #1]
Pair the following atoms (your labels will differ!)
N1t with C3
N2t with C2
C17t with C16
Be careful with matching the correct orientation of the molecule, interchanging of N1t and
N2t would change it
For "Maximal coincidence distance" type 0.8
Press "Show coinciding atoms"
The expected coincidence ratio in our example is 14/22 (depends which carbon atoms were
previously localized from the difference Fourier)

Press "Apply+End"
Start "Edit atoms"
Delete all carbon atoms, which are not blue (i.e. carbons out of molecules).
Be careful not to delete chlorine
Press "Select all"
The structure should contain 36 atoms
Press "Refresh"; OK; YES to save changes
Plot the structure using "Draw average structure

305 Jana2006 Cookbook, version October 2017

 Start :File → Structure → Save As" and make a backup copy

15. Preparing refinement with fixed rigid bodies

Start "Edit atoms"
Select carbons of phenyl, which have hydrogen: C2p, C3p, C4p, C5p and C6p
Start "Action → Adding of hydrogen atoms"
Check "generate keep command for Refine"
Check "try automatic run ..."
OK
Repeat procedure for carbon atoms of phenantroline: C7t, C8t, C9t, C11t, C12t, C14t, C15t
and C16t
The procedure for adding hydrogen atoms, applied to already existing hydrogens, generates
keep commands, which are necessary for subsequent refinement.
After this process number of atoms in the structure must be unchanged, i.e. 36
Start Refinement options; go to the page "Various"; press "Fixed commands"
Select "All parameters"
For "Atoms/parameters" type "C*p C*t N*t"; Add; OK; OK; YES to save commands without
starting Refinement
This fixes geometry of both phenyl molecules and phenantroline

16. Refinement with rigid bodies

Start "File → Structure → Save As" and make a backup copy
Start Refinement options; go to the page Basic
Clear "Disable atoms having too large isotropic ADP parameter"
OK; YES+START
Rall will be ~28% for 61 parameters
Start "Parameters → Molecules → Edit"
Set 1 position modulation wave for the first position of Ph
Set 1 position modulation wave for the second position of Ph
Set 1 position modulation wave for the third position of Ph
Set 1 position modulation wave for the first position of Trol

306 Jana2006 Cookbook, version October 2017

Run refinement
Rall will be ~22% for 109 parameters
Start "Edit atoms"
For Sn1 set harmonic ADP
For Cl1 and O1 set 2 position modulation waves and harmonic ADP
Run refinement
Rall will be ~20% for 136 parameters
Start Refinement options; go to the page "Various"; press "Fixed commands"
Disable the command "fixed all C*p C*t N*t "
OK; OK; YES+START
The refinement converges with Rall~19%, Rmain~11%, Rsat1~22% and Rsat2~27% for 204
parameters
Start "Edit atoms"
Set harmonic ADP for carbons and nitrogens of both molecules
OK; OK, YES to save changes
Start "Parameters → Molecules → Edit"
Set 2 position modulation wave for the first position of Ph
Set 2 position modulation wave for the second position of Ph
Set 2 position modulation wave for the third position of Ph
Set 2 position modulation wave for the first position of Trol
Run refinement
The refinement converges with Rall~9%, Rmain~6%, Rsat1~11% and Rsat2~10% for 352
parameters
Start "Edit atoms"
Set one ADP modulation wave for Sn1, Cl1 and O1
Start "Parameters → Molecules → Edit"
Set one ADP modulation wave (modulation of TLS parameters) for three positions of Ph and
one position of Trol
Run refinement
The refinement converges with Rall~8%, Rmain~6%, Rsat1~10% and Rsat2~10% for 548
parameters
307 Jana2006 Cookbook, version October 2017
 Start "Edit atoms"
Set another one ADP modulation wave for Sn1, Cl1 and O1
Start "Parameters → Molecules → Edit"
Set another one ADP modulation wave for three positions of Ph and one position of Trol
Run refinement
The refinement converges with Rall~8%, Rmain~6%, Rsat1~10% and Rsat2~9% for 744
parameters
This is the final structure model with molecules

17. Analysis of results

Considering very good data quality, the results are still not satisfactory. Here we are looking
for problems in the structure model
Start "Edit/View → View of Refine"
There are no serious warnings
Plot the supercell structure
The program does not use the transformation to five-fold supercell. Instead, it expands the
structure to 5x1x5 supercell.
Geometry seems to be correct.
Calculate difference Fourier map
YES to start procedure for including new atoms
Do not include any maxima but look at their distances to atoms of our structure model. The
largest difference maxima are close to atoms of phenantroline and phenyl, indicating some
problem with the rigid body description.
Start options for “Dist”
Press "Planes"
For "Atoms defining the best plane" type: C1pa C2pa C3pa C4pa C5pa C6pa; Add
For "Atoms defining the best plane" type: C1pb C2pb C3pb C4pb C5pb C6pb; Add
For "Atoms defining the best plane" type: C1pc C2pc C3pc C4pc C5pc C6pc; Add
The second and third line can be created with help of the button "Clone"
OK
Go to the page "Modulation"
Check "Make coordination t-tables"
OK, YES+START
Start "Edit/View → View of Dist"
"Go to" → "List dihedral angles"

Each "t" value indicates one of the defined commensurate sections, which is realized as one
configuration in the five-fold superstructure. We can see that changes of angles between the
planes #1, #2 and #2, #3 reach 30 degrees, which cannot be reliably described by rigid body
modulation function as used in Jana2006.
308 Jana2006 Cookbook, version October 2017
 Start “Grapht”
Press “New/edit”
For “Parameter to draw” set “plane”
For “p” set limits from 0 to 90
For “Plane #1” type “C1pa C2pa C3pa C4pa C5pa C6pa”
For Plane #2” type “0 0 1”
Grapht now displays angle between the plane through the first position of Ph and the axis c.
Press “Options”; check “Indicate commensurate t sections”; OK
Grapht now also displays the sections, for which the structure is defined. The orientation of
the single plane changes by ~40 degrees. Again, this cannot be reliably described by rigid
body modulation function as used in Jana2006.

Actual value of t-sections is connected with the origin selection made automatically during
solution by Superflip. Also the phenyl molecules can be placed in different order compared
with this manual. Thus the actual plot will differ.

18. Creating structure model without phenyl molecules

Start "File → Structure → Save As" and make a backup copy
Start "Edit atoms"; Select atoms *p
Start "Action → Atoms from molecule to atomic part"
The three molecular position have been transformed to atoms with suffixes "a", "b" and "c"
Select H*pa, H*pb and H*pc (it can be also done with H*p?)
Start "Action → Edit/Define"
Set isotropic ADP
OK; OK; YES to save changes
Start Refinement options
Go to the page "Various"; press "Keep commands"; press "Edit"
Using the text editor:
- change labels *p to *pa
- duplicate the commands and change *pa to *pb
- create the same commands for *pc
OK
Go to the page "Basic"; set Damping factor 0.5
OK; YES+START
Run refinement

309 Jana2006 Cookbook, version October 2017

 The refinement converges with Rall~5%, Rmain~4%, Rsat1~6% and Rsat2~6% for 1176
parameters.

19. Creating structure without phenantroline molecules

Start "Edit atoms"; Select atoms *t
Start "Action → Atoms from molecule to atomic part"
The molecular position have been transformed to atomic part without adding suffix "a",
because the molecule had only one position. For this reason, keep commands are still valid
and no editing is needed.
Select H*t
Start "Action → Edit/Define"
Set isotropic ADP
OK; OK; YES to save changes
Run refinement
The refinement converges with Rall~2.7%, Rmain~2.3%, Rsat1~3.2% and Rsat2~2.9% for 1576
parameters

20. Geometry of the atomic model

The atomic model has excellent R factors. However, we have to ensure that the phenyls are
planar and not distorted.
Start "File → Structure → Save As" and make a backup copy
Start Refinement options
Go to the page "Various"; press "Keep commands"
Select "Geometry" and "Plane"
For "List of atoms" type C1pa C2pa C3pa C4pa C5pa C6pa; Add
For "List of atoms" type C1pb C2pb C3pb C4pb C5pa C6pb; Add
For "List of atoms" type C1pc C2pc C3pc C4pc C5pc C6pc; Add
OK; OK; YES+START
Change in R values should be very small
Start Dist
Commands for best planes prepared previously are still valid
Start "Edit/View → View of Dist"
"Go to" → "List of best planes"
Here we can see that the phenyls are almost exactly planar, as expected
Start Grapht
Press "New/Edit"
For "Parameter to draw" select "ADP parameter"
For "p" define "0" as Minimum and "0.1" as Maximum
For type of parameter to plot, select "U(max)"
For Central atom type "N1t"
OK
[On the screen: plot of U(max) for N1t]
Press "Options"
Select "Indicate commensurate t sections"; OK
Now the plot shows the section where the structure is defined
Press "New/Edit"
Press "New Item"; for "Central atom" type "N2t"
Press "New Item"; for "Central atom" type "C7t"
310 Jana2006 Cookbook, version October 2017
Press "New Item"; for "Central atom" type "C8t"
Add New items for C8t - C18t
OK
[On the screen: cumulative plot of U(max) for all atoms of the phenantroline]

The plot indicates that for one of the sections (here for t=0; its actual value is connected with
the origin selection made automatically during solution by Superflip) some of the
phenantroline atoms are strongly disordered, while the other not. This could not be
described with the rigid body model.
Quit Grapht
Draw the average structure, select only the phenantroline molecule
For this Ctrl-M keyboard shortcut of Diamond can be used, which selects molecule belonging
to a selected atom
Note atoms N1t, C18t, C17t and N2t.
Phenantroline is often twisted and we will calculate torsion angle of these atoms

Start Dist options

Press "Torsion angles"
For atoms type N1t, C18t, C17t and N2t; Add
OK; OK; YES+START
Open listing of Dist
"Go to" → "List of torsion angles"

311 Jana2006 Cookbook, version October 2017

 Here we can see that for five commensurate sections the torsion angle fluctuates between -
4 and +12 degrees. This torsion could not be described by the molecular model.

The same torsion angle can be visualized in Grapht:

(Note that only the indicated sections are defined)

21. Water
The last missing atoms in our structure model are hydrogens of the water molecule
Start Fourier options
For "Map type" select difference Fourier
Check "Apply sin(th)/lambda limits
define "max" as 0.5
OK; YES+START
YES to start procedure for including new atoms
Add two difference maxima (probably the first two ones) 0.82Å close to O1 and name them
H1O1 and H2O1
Start "Parameters → Molecules → New molecule"
Select "Atomic part" and "water" as a name; Next
Select O1, H1O1 and H2O1; OK
For the reference point select "Explicit" and use O1; NEXT
Apply+End
Start "Edit atoms"
Delete all modulation waves for O1, H1O1 and H2O1; OK; YES to save changes
Start "Parameters → Molecules → Edit"
Set one position modulation wave for "water"; OK; YES to save changes
Start Refinement options
In page "Various" define distance restraint 0.82Å for O1-H1O1 and O1-H2O2
In page "Various" define keep command for ADP of hydrogens H1O1, H2O1, to be 1.2
multiple of Ueq(O1).

312 Jana2006 Cookbook, version October 2017

 Run refinement
Add another position modulation wave for the molecule "water"
Run refinement
In "Edit atoms" set one ADP modulation wave for O1
Run refinement
The refinement converges with Rall~2.7%, Rmain~2.2%, Rsat1~3.0% and Rsat2~2.8% for 1584
parameters
The first ADP modulation wave improved only slightly R values. Therefore, we will not add
the second wave - this is the principle of saving parameters in commensurate refinement
comparing with refinement in the supercell.
In the final step introduce isotropic extinction
Run refinement
The refinement converges with Rall~2.6%, Rmain~2.1%, Rsat1~3.0% and Rsat2~2.8% for 1585
parameters
This is the final commensurate model

22. Transformation to supercell

The final commensurate model can be transformed to the supercell
Start "Tools → Transformations → Go to supercell structure"
For Name type "sup"; OK
OK to transform molecules
Follow steps of the data processing wizard
YES to continue with the new structure
Set zero refinement cycles
Run refinement (skip warnings)
The R value will be 2.6%, because both descriptions are equivalent
In order to refine the supercell structure, keep commands for hydrogens must be re-created.

The supercell structure contains 15 independent phenyls:

C1pa-1, C2pa-1, C3pa-1, C4pa-1, C5pa-1, C6pa-1
....
C1pa-5, C2pa-5, C3pa-5, C4pa-5, C5pa-5, C6pa-5
C1pb-1, C2pb-1, C3pb-1, C4pb-1, C5pb-1, C6pb-1
....
C1pb-5, C2pb-5, C3pb-5, C4pb-5, C5pb-5, C6pb-5
C1pc-1, C2pc-1, C3pc-1, C4pc-1, C5pc-1, C6pc-1
....
C1pc-5, C2pc-5, C3pc-5, C4pc-5, C5pc-5, C6pc-5

Five independent phenantroline molecules:

N1t-1, N2t-1, C7t-1, C8t-1, C9t-1, C10t-1, C11t-1, C12t-1,
C13t-1, ..., C17t-1, C18t-1
....
N1t-5, N2t-5, C7t-5, C8t-5, C9t-5, C10t-5, C11t-5, C12t-5,
C13t-5, ..., C17t-5, C18t-5

Five independent water molecules:

O1a-1; O1a-2; O1a-3; Oa-4; O1a-5

313 Jana2006 Cookbook, version October 2017

 Keep commands can be easily expanded and edited according to the above mentioned
labels, however, they can be also copied from the pre-prepared file supercell_keep.txt.
Start Refinement options
Go to the page "Various"; press "Keep commands"
Press "Edit"
Replace content of the file with content of the supercell_keep.txt
Save and close the text editor
OK
Number of keep commands should be 235
Press "Distance restraints"
Restrain the following distances to 0.82Å: O1a-1 H1O1a-1; O1a-1 H2O1a-1; O1a-2 H1O1a-2;
O1a-2 H2O1a-2; O1a-3 H1O1a-3; O1a-3 H2O1a-3; O1a-4 H1O1a-4; O1a-4 H2O1a-4; O1a-5
H1O1a-5; O1a-5 H2O1a-5;
Some keep commands and distance restraints may be invalid because the atomic positions
may be transformed during the commensurate-to-supercell transformation. Such commands
would need symmetry codes.
Start "Edit atoms"
Press "Select all"
Start "Action" → "create symmetrically contiguous motifs"
After this the keep commands can use labels without symmetry codes
Start refinement (still with zero cycles)
The R value will be 2.6%
Start "File → Structure → Save As" and make a backup copy
Set 100 refinement cycles, damping factor 0.5, and run refinement
Refinement uses 1611 parameters, which is more than in comensurate refinement, because
we did not use full set of parameters for water molecules. However, the resulting R value is
still 2.58%

23. Back to the rigid body description

Now we can introduce phenyl molecule in 15 positions, which was not possible in the
commensurate refinement due to too large rotations. For phenantroline, we can try to use
rigid body description using two halves of the phenantroline as independent pattern
molecules, in order to describe the torsion of phennatroline. Finally, we can model the
disordered phenantroline molecule by partially occupied molecules. All this is routinne job,
which will not be presented here.

24. Conclusions
Commensurate refinement can help to find solution of very large structures, it allows to save
parameters by using less than the full set of modulation parameters, and it facilitates
interpretation of a superstructure through calculation of various parameters as a function of
t. However, commensurate refinement may be difficult when satellites are too strong,
because strong satellites indicate large displacements, which must be described by
modulation functions defined only in few points. Despite of this, commensurate refinement
was possible for our structure and it helped us to understand its geometry. On the other
hand, the phenyl rigid body could be only possible in the supercell refinement, as well as the
full description of the disordered phenantroline molecule.

314 Jana2006 Cookbook, version October 2017

Example 8.1: TerBut
Application of rigid body groups for disordered parts of crystal structure
Revised: 4 March 2015

N-(3-nitrobenzoyl)-N’,N’’-bis(tert-butyl)phosphoric triamide
Chemical formula: C90H150N24O24P6

Single crystal data measured with Oxford Diffraction four-circle diffractometer

Input files: TerBut.hkl, TerBut.cif_od
Frame scaling, absorption correction: done with software of the diffractometer

1. Creating new jobname

Start Jana2006
“File → Structure → New” opens a file manager
Left pane: locate directory with input files
Right pane: double-click TerBut

2. Import Wizard
Select “Known diffractometer formats”; NEXT
Select “Oxford Diffraction – CCD”; NEXT
Leave all settings unchanged; NEXT
Leave all settings unchanged; NEXT
For absorption correction select “None or done before importing”; NEXT
FINISH

3. Data Repository
Each line in data repository window corresponds to one data set.
OK; YES to accept the data set

4. Symmetry Wizard
[On the screen: “Information” about symmetry wizard]

315 Jana2006 Cookbook, version October 2017

 NEXT
[On the screen: “Tolerances for crystal system recognition”.]
Leave all settings default; NEXT
[On the screen: Select Laue symmetry]
Select Triclinic; NEXT
Select primitive unit cell; NEXT
[On the screen: Select space group]
Select P-1; NEXT
Accept the space group; FINISH

5. Creating refinement reflection file

NEXT to confirm threshold 3sigma
[On the screen: settings for merging of reflections]
Leave all settings default; NEXT
Press "Accept the new DatBlock and calculate coverage"; FINISH

6. Structure Solution Wizard

[On the screen: "Information" about Structure solution tool]
OK
[On the screen: window of Structure solution]
In “Formula” textbox type list of chemical elements for TerBut: C15 N4 O4 P H25;
Set "Formula units" to 2 and calculate density;
[On the screen: Calculated density: 0.4242 g/cm3]

This is unrealistic value which indicates that the structure can contain more independent
structural motifs. As the density should be larger than 1 the most probable number of
independent structural motifs is 6.
Set "Formula units" to 6;
Select “Superflip”, “Peaks from Jana2006”; leave other settings default; “Run solution”
Close listing of Superflip
Press Accept last solution” to leave the structure solution wizard

7. Verification of the structure solution

Start “Plot structure”, use either "Draw+return" or "Draw+continue"
[On the screen: window of Diamond]
Start “Build → Get molecules” or use on the bottom toolbar
Rotate the molecule ( and on the bottom toolbar) to get an optimal view
[On the screen: plot of the three independent, probably incomplete molecules]
Assignment of chemical types need not be correct and some atoms may be missing.
Chemical types will be corrected later on.
The three iindependent molecules are combined with the inversion center. This confirms
Z=6 as estimated previously from the density. Correct Z helps the program to offer proper
number of maxima from difference Fourier.
Quit Diamond;

8. Refinement
[On the screen: basic window of Jana]
Right-click the icon of Refine.
316 Jana2006 Cookbook, version October 2017
 [On the screen: Refinement commands]
Define 100 of cycles; OK
Choose “YES+START”
Refinement converges with R(obs) value about 20-22%

9. Completing the structure – difference Fourier synthesis

[On the screen: basic window of Jana]
Right-click the icon of Fourier.
[On the screen: Fourier commands]
For "Map type" select difference Fourier, make sure that “Weighting of reflections” is
checked and “Apply sin(th)/lambda limits” unchecked
OK; YES+START to save commands and run the difference Fourier;
YES to start the procedure for including of new atoms;
[On the screen: parameters for inserting/replacing of atoms]
Use default parameters; OK
[On the screen: the form for inserting of new atoms]
Go through the list of maxima and include all peaks having reasonable distances (i.e. similar
to C-C distance) to already localized atoms. Skip short distances. Do not attempt including H
atoms.
After pressing the button “Include selected peak” the form for completing of the new atom
appears. For the “Name of the atom” type C*. For more information what C* means press
the “Info” button.
Repeat the points 8 and 9 till no new atoms are localized. The structure should have 72 non-
H atoms
FINISH; YES to include new atoms
Repeat refinement
Finally R(obs) value should be between 16 and 17%

10. New verification of the structure solution

Start “Plot structure”, "Draw+continue" or "Draw+return"
[On the screen: window of Diamond]
Start “Build → Get molecules” or use on the bottom toolbar
Rotate the molecule ( and on the bottom toolbar) to get an optimal view
[On the screen: plot of the three independent molecules]
Activate atom labels
[On the screen: plot of the molecule with atom labels]
Compare the plot with the chemical scheme at the beginning of this Cookbook chapter and
note labels of atoms which chemical types need to be changed.
(For Diamond started by "Draw + return" the changes can be done immediately in Jana2006
running alongside.)
Use at the bottom toolbar of Diamond to see temperature ellipsoids.
The large spheres derived from Uiso values for the terminal carbon atoms indicate positional
disorder of some tertiary butyl groups.

11. Assignment of correct chemical types

[On the screen: basic Jana window]

317 Jana2006 Cookbook, version October 2017

 Start “Edit atoms”
[On the screen: list of atoms]
Highlight one of the atoms you are going to modify
Make left double-click or Start “Action → Edit/define”
Change “Type” to the desired one
Rename the atom. Here you can use again the wild character * which prevents duplication
of atom names.
Repeat this procedure till all atoms have correct chemical types.
Note: alternatively you can select more atoms and change the chemical type of the selected
group. After returning to the list of atoms in Atom/Edit the atoms are still selected and they
can be automatically renamed with “Action → Rename selected atoms”.
Check again the chemical types with Diamond
Check number of atoms with "Edit atoms", using the textbox close to "Select advanced",
which accepts wildcards as "C*". The structure contains 3P, 12O, 12N and 45C; 72 atoms
altogether.

12. Refinement
[On the screen: basic window of Jana]
Double-click the icon Refine.
Refinement converges with R(obs) value about 16%
Start “Edit atoms”
[On the screen: list of atoms]
Press “Select all” and “Action → Edit/define”
Choose “harmonic” for ADP parameters. This defines anisotropic temperature parameters
for all atoms. OK, OK, YES to save changes
Double-click the icon Refine.
Refinement converges with R(obs) value about 9%

13. Formula check

[On the screen: basic window of Jana]
Run EditM50;
Go to the tab “Composition” and press button “Formula from M40”.
The derived formula should be correct for all atom types except hydrogen.

14. Unifying atom labels

To follow easily the cookbook we need to re-label the atoms. It would be tedious work and
therefore we will used pre-prepared files with correct labelling.
Start “File → Structure → Copy in”, select the structure “TerBut-simple-atomic” and in the
next dialogue select files m40 and m50 for the copy.
Double-click the icon Refine.
Refinement should converge to R(obs) value about 9.4%
The labeling of atoms of the first independent molecule is shown in the following figure:

318 Jana2006 Cookbook, version October 2017

 The last digit in each label is the identifier of the molecule (replaced by 2 and 3 for the
second and third independent molecule, respectively).

15. Search for split atom positions

Double-click the icon Contour; Click the button “New plot”
[On the screen: options for defining of new maps]
In “New map to draw” select “Fourier map”;
Select “Calculate new ones” and “Draw a general section”; OK
[On the screen: Fourier commands – page Basic]
Set “Map type” to “F(obs) – Fourier”;
Clear “Omit not-matching reflections”;
Clear “Apply sin(th)/lamba limits”;
Select “Use weighting of reflections”; OK
[On the screen: Plane/Volume]
For 1st, 2nd and 3rd atom use C91, C101 and C111;
Keep interpolation step 0.1 Å;
Define scope of the section: 5 5 0 and “1st point to” 1.5 1.5 0; OK
You should see the following figure:
e
3
0
0
0
0
3
.
5

e
2

2
.
5

1
.
5

0
.
5

-
0
.
5

-
1
.
5
e
2
5
0
.
5
-
0
.
5
-
1
.
5

319 Jana2006 Cookbook, version October 2017

Click the button “New plot”
(The definition of the plane has been automatically saved as Plane#1)
[On the screen: options for defining of new maps]
Select “Use old maps” and “Draw a general section” (i.e. default values); OK
[On the screen: Plane/Volume]
Select “Use old”
Rewrite 1st, 2nd and 3rd atom by C92, C102 and C112;
You should see similar figure like for the previous figure
Repeat the procedure for C93, C103 and C113;
Again no clear splitting of atomic positions
Repeat the procedure for C131, C141 and C151.
The last map indicates clearly splitting for two atoms:

Locate the split positions by clicking at corresponding points of the section and save them
for future use by “Save the selected point for future use” button.

points close to C141 ~ (0.2795, 1.102,0.2494) and (0.2995,1.1158,0.1820)

(they can be named C141a and C141b)
points close to C151 ~ (0.2455,1.111,0.1117) and (0.1869,1.0974,0.1017)
(they can be named C151a and C151b)

Make the same for the second molecule (the section C132, C142, C152). Here the splitting
of atoms C142 and C152 is less pronounced, however we only need some points for initial
orientation of rigid bodies.

320 Jana2006 Cookbook, version October 2017

points close to C142 ~ (0.7732, 0.3731,0.4747) and (0.8239,0.4088,0.4588)
(they can be named C142a and C142b)
points close to C152 ~ (0.8695,0.4942,0.5074) and (0.8356,0.5156,0.5715)
(they can be named C152a and C152b)

Make the same for the third molecule (the section C133, C143, C153). Here all positions are
clearly split but we only need two pairs of points..

points close to C143 ~ (-0.1854, 0.7837,0.1733) and (-0.1759,0.7836,0.1320)

(they can be named C143a and C143b)
points close to C153 ~ (-0.0588,0.8945,0.1860) and (-0.1022,0.8639,0.2186)
(they can be named C153a and C153b)

The split positions localized during preparation of this example were also written to the file
“TerBut-points.txt” and can be later to pasted to Jana2006 textboxes.

321 Jana2006 Cookbook, version October 2017

 Quit Contour
YES to save new points in M40 (this does not change structure model)
The points have been stored in M40 in the “Saved points” section

16. Defining of rigid-body groups

First we will create a model where one pattern “molecule” – tertiary butyl – is used to create
three non-split positions related to N3i, C8i, C9i, C10i, C11i fragments and another six split
positions related to N4i, C12i, C13i, C14i and C15i fragments (i=1,2,3).
Start “File → Structure → Save As” and create a backup copy “before_rigid” but do not
continue with the new structure.
Start “Parameters → Molecules → New molecule”;
[On the screen: Wizard for new molecule]
Select “Atoms of the new molecule from” - "Atomic part";
For “Name of the molecule” type, for instance, "Tb"; NEXT;
[On the screen: Select atoms for the molecule]
Select atoms to be used as a pattern (elsewhere called model) molecule – N31, C81, C91,
C101 and C111;OK
[On the screen: Define the molecular reference point]
Select “the molecule reference point” as “Explicit” defined by the atom C81; NEXT
[On the screen: Form for completing of the molecular position #1];
The “Coincidence ratio” says that five atoms of the molecular position #1 coincide with five
atoms in the atomic part of M40. This should be true because the pattern molecule is
created from five atoms in M40 and the first position coincides with the pattern molecule.
Use the default value for occupancy; Press “Apply+Next Position”;
[On the screen: Form for completing of the molecular position #2];
For the position #2 and all consecutive ones we have to define where to put the pattern
molecule to get the new actual position. This means that actual positions of three atoms of
the pattern molecule, not lying in a line, are to be specified.
Fill “Model atoms” column with N31, C101 and C111 and the fill the “Actual position/atom”
column with N32, C102 and C112. Press the button “Calculate molecular parameters”.
The coincidence ratio should be 5/5. It says that five atoms of the pattern molecule coincide
in its actual position with five already existing atoms of the structure. After finishing the
molecular position the coinciding atoms will be deleted from the atomic part – for this
reason it is important that all of them are recognized by the program. The coincidence can
be controlled by the "Maximal coincidence distance".
Press the button “Apply+Next Position”;
Repeat the procedure for the third non-split tertiary group (i.e. N33, C103 and C113 for
Actual positions); press "Apply+Next position"
[On the screen: Form for completing of the molecular position #4];
For actual positions use N41, % C141a and %C151a. The character "%" refers to saved
points.
Press the button “Calculate molecular parameters”.
If the "Coincidence ratio" is not 5/5, change the “Maximal coincidence distance” to 0.8 and
press again "Calculate coincidence distance"
If the "Coincidence ratio" is still not 5/5, replace the points for the 2nd and 3rd actual
position and press again "Calculate coincidence distance"
The reason for this step is that the selected saved points may define by mistake
contradicting rotations of the pattern molecule. This cannot happen when the "a" and "b"
322 Jana2006 Cookbook, version October 2017
 labels have been used as indicated in the plots.
Finally, the “Coincidence ratio” should be 5/5
For “Occupancy" type 0.5.
Press “Apply+Next position”;
[On the screen: Form for completing of the molecular position #5];
For actual positions use the same atom N41 and another two split positions % C141b and
%C151b.
Press the button “Calculate molecular parameters” and check that the "Coincidence ratio"
is 5/5;
Change the occupancy to 0.5;
Press the button “Apply+Next Position”;
Repeat the process for the second and third split tertiary butyl:
Position #6: N42, %C142a, %C152a
“Apply+Next Position”
Position #7: N42, %C142b, %C152b
“Apply+Next Position”
Position #8: N43, %C143a, %C153a
“Apply+Next Position”
Position #9: N43, %C143b, %C153b
“Apply+End”
The total number of molecular positions is 9.

17. Verification
Use Diamond for verification of proper geometry: each molecule should have one single
and one disordered tertiary group.
Use Parameters → Molecules → Edit for verification that positions #4 – #9 have occupancy
(i.e. the parameter “aimol”) 0.5
Use “EditM50”, page “Composition”, button “Formula from M40”
If occupancy of molecular positions has been properly defined and the coincidence ratio has
been always 5/5 the formula should be C15 N4 O4 P H0 (for Z=6)
Use “Edit atoms” to see changes in the atomic part of M40
Black atoms are the ones not used in molecular description
Blue atoms are the ones defining the pattern molecule
Start “File → Structure → Save As” and create a backup copy “rigid_ini” but do not
continue with the new structure.

18. Refinement
[On the screen: basic window of Jana]
Right-click the icon of Refine.
[On the screen: Refinement commands]
Change the damping factor to 0.5;
OK, “YES+START”
Refinement converges with R(obs) value about 10.4%

19. Refinement of the molecular occupancies

Start “Parameters → Molecules → Edit”;
[On the screen: molecule edit]
Allow refinement of the molecular occupancies (aimol) for the positions #4, #6 and #8;

323 Jana2006 Cookbook, version October 2017

 OK; YES to save changes
Right-click the icon of Refine; select tab "Various”;
[On the screen: Refinement commands]
Press the button “Restrictions”;
[On the screen: The tool for defining and editing of restriction commands]
Write into the textbox “Atoms/molecules”: “Tb#4 Tb#5”
Deselect “Coordinates”
Select “Occupancies: Keep overall sum”
Select “ADP parameters”
Press the button “Add”;
This restriction means that the molecular positions #4 and #5 have independent coordinates,
the same ADP and complementary occupation
Select the line: restric Tb#4 Tb#5;
Press the button “Clone” and then “Load” and modify the string to “Tb#6 Tb#7”
Press the button “Rewrite”;
Press the button “Clone” and then “Load” and modify the string to “Tb#8 Tb#9”
Press the button “Rewrite”;
OK; OK; Yes+Start;
Refinement converges with R(obs) value about 10.3% with 475 parameters

20. Refinement of the molecular TLS tensors

Start “Edit atoms”;
[On the screen: tool for selection of atoms]
Select all atoms of the pattern molecule (before selection they are indicated by blue color).
Double-click; Select “Use TLS”;
OK; OK; Rewrite changes
This change means that individual ADP parameters of the pattern molecule are suppressed.
Instead, for each actual position independent TLS tensors will be refined.
Run refinement
Refinement converges with R(obs) value about 8.6% with 565 parameters

21. Adding hydrogen atoms

Start “Edit atoms”;
[On the screen: tool for selection of atoms]
Select all carbon atoms and nitrogen atoms N21, N22 and N23
Press the button “Action”
Select the procedure “Adding of hydrogen atoms”
[On the screen: parameters for adding hydrogen atom]
Use default options (i.e. automatic procedure for C and manual procedure for N);
For N21, N22 and N32 select the trigonal coordination with two neighbors (P and C). Then
only one H atom will be added.
Total number of hydrogen atoms in Atoms edit should be 24
OK; Save changes
Use Diamond for verification that all hydrogen atoms have been correctly placed
Run refinement
Refinement converges with R(obs) value about 6%

324 Jana2006 Cookbook, version October 2017

22. Drawing of the refined model
Start "Plot structure"
In Diamond, start “Build → Get molecules” or use on the bottom toolbar
Use at the bottom toolbar of Diamond to see temperature ellipsoids. For plotting H
atoms without ellipsoids switch to “Table of atom groups”, select all H atoms and use
again.
The large temperature ellipsoids for the terminal carbon atoms indicate that there is
probably some additional positional splitting of tertiary butyls N3, C8, C9, C10, C11 and also
of some oxygen atoms. For teaching purposes, we can stop at this stage, as introducing of
additional splitting does not bring any new.
Quit Diamond
Start “File → Structure → Save as” and save the result as “TerBut_final_1” but do not
continue with the new structure.

23. Reset the structure into the atomic model

Start “File → Structure → Copy in”, select the structure “before_rigid” and copy in the m40
and m50 file.

24. Defining of rigid-body groups from a known model saved in M45

M45 may contain coordinates of a rigid-body group. Syntax of M45 is explained in the
example 2.3.2-MagCag, where steps how to create M45 from a template found in CSD are
shown. Library of frequently used rigid bodies for Jana2006 is under development, see the
directory "Molecular patterns" in Jana2006 installation path.
Start “Parameters → Molecules → New molecule”;
[On the screen: Wizard for new molecule]
Select “the new molecule from a model file”;
Define “Name of the molecule” Tb;
Define “Model filename” TerButyl-no-symmetry.m45;
Keep the default scaling distance
With this number the whole object can be shrunk or expanded
Press "Show the model molecule"
This will display the pre-prepared figure from the file TerButyl-no-symmetry.jpg
NEXT
[On the screen: specify atomic types in the model molecule]
Keep the default atomic types; NEXT
[On the screen: Define the molecular reference point]
Select “the molecular reference point” as explicit defined by the atom C81; NEXT
[On the screen: Form for completing of the molecular position #1];
For creating molecular parameters from a model file we have to specify also the actual
coordinates of the first position. Starting from the position #2 we can repeat instructions
from the point 16 and 17. Because the pattern molecule contains hydrogen atoms which are
not present in the atomic part, the coincidence ratio will be 5/14 or higher for the split
positions. This is due to the fact that some hydrogen atoms of the split positions can be close
to average carbon positions. This should not have any effect as this message has only
informative character.For coincidence lower than 5 please change the "Maximal coincidence
distance" to 0.8.

325 Jana2006 Cookbook, version October 2017

 As follows from the point 16 and 17, the atoms N31, C101 and C111 of the pattern molecule
will be associated with the following actual positions present in the atomic part of M40 or
from the saved points:

1. N31 C101 C111 (occupancy 1)

2. N32 C102 C112 (occupancy 1)
3. N33 C103 C113 (occupancy 1)
4. N41 %C141a %C151a (occupancy 0.5)
5. N41 %C141b %C151b (occupancy 0.5)
6. N42 %C142a %C152a (occupancy 0.5)
7. N42 %C142b %C152b (occupancy 0.5)
8. N43 %C143a %C153a (occupancy 0.5)C101
9. N43 %C143b %C153b (occupancy 0.5)

Resulting number of atoms: 42 in atomic part (here hydrogens are still missing); 14 in the
pattern molecule.

25. Refining of the structure with fixed ideal geometry

Start “Parameters → Molecules → Edit”;
[On the screen: molecule edit]
Allow refinement of the molecular occupancies (aimol) for the positions #4, #6 and #8;
OK; YES to rewrite files;
Right-click the icon of Refine;
[On the screen: Refinement commands]
Change number of cycles to 100 and the damping factor to 0.5;
Select tab “Various”;
[On the screen: Refinement commands]
Press the button “Restrictions”;
[On the screen: The tool for defining and editing of restriction commands]
Write into the textbox “Atoms/molecules”: “Tb#4 Tb#5”
Deselect “Make identical coordinates”;
Select: “Occupancies: Keep overall sum”
Select: “ADP parameters”
Press the button “Add”;
This restriction means that the molecular positions #4 and #5 have independent coordinates,
the same ADP and complementary occupation
Select the line: restric Tb#4 Tb#5;
Press the button “Clone” and then “Load” and modify the string to “Tb#6 Tb#7”
Press the button “Rewrite”;
Press the button “Clone” and then “Load” and modify the string to “Tb#8 Tb#9”
Press the button “Rewrite”; OK
[On the screen: Refinement commands]
Press the button “Fixed commands”;
Select “Coordinates”; Fill textbox “Atoms/parameters” with
N31 C81 C91 C101 C111 H?c91 H?c101 H?c111
This fixes coordinates of the pattern molecule. For library molecule we expect correct
geometry.
Press the button “Add”; OK
326 Jana2006 Cookbook, version October 2017
 [On the screen: Refinement commands]
Select “Keep commands”;
Select “ADP”, “Riding”;
Central atom: C91; Number of hydrogen atom 3;
Fill hydrogen textboxes by: H1C91, H2C91 and H3C91;
For this button “Select hydrogens” can be used.
Add;
Repeat the process for C101 and C111;
OK; OK; Yes+Start
Refinement converges with R(obs) value about 10.2% for 441 parameters.

26. Refining of the structure with anisotropic ADP for tertiary butyls
Start “Edit Atoms”;
[On the screen: tool for selection of atoms]
Select non-hydrogen atoms of the pattern molecule.
Double-click; Select "Harmonic ADP";
OK; OK; Rewrite changes
Start refinement;
Refinement converges with R(obs) value about 9.4% for 466 parameters

27. Refining of the structure with TLS tensors for tertiary butyls
Start “Edit Atoms”;
[On the screen: tool for selection of atoms]
Select all atoms of the pattern molecule.
Select “Use TLS” using double-click;
OK; OK; Rewrite changes
Start refinement;
Refinement converges with R(obs) value about 7.5% for 556 parameters

28. Adding remaining hydrogen atoms

Start “Edit Atoms”;
[On the screen: tool for selection of atoms]
Select all individual (black) carbon atoms and nitrogen atoms N21, N22 and N23
It is important that carbon atoms of the molecule are not selected. Here the geometric keep
commands would contradict with setting of the local symmetry, which will be done later on
Press the button “Action”
Select the procedure “Adding of hydrogen atoms”
[On the screen: parameters for adding hydrogen atom]
Use default (i.e. automatic run);
For N21, N22 and N32 select the trigonal coordination with the neighbors (P1, C71), (P2
C71) and (P3, C73), respectively.
OK; Save changes
Repeat refinement
Refinement converges with R(obs) value about 6.6% for 556 parameters. This result can be
compared with the previously obtained model TerBut_final_1, where the shape of the
pattern molecule was refined, yielding R value 5.8 for 565 parameters.
Start “File → Structure → Save as” and save the result as “TerBut_final_2” but do not
continue with the new structure.

327 Jana2006 Cookbook, version October 2017

29. Introducing of the local symmetry
The tertiary butyl has in an ideal case the point symmetry 31m . Application of the local
symmetry can considerably reduce the number of independent atoms. The point group
generated by Jana2006 has its three-fold axis directed along the z-axis and the normal of the
mirror plane directed along x-axis. As the atoms N31 and C81 are localized at the three-fold
axis of the molecule the vector from the central point C81 to N31 can be used to define the z
axis. The second x-axis is to be perpendicular to the z axis and it can be located at the plane
N31-C81-C91.
Start “Parameters → Molecules → Edit”;
Define the point group 31m; TAB
A dialog window for defining the local coordinate system for the pattern molecule will
appear.
The model imported from the m45 file fulfills already the 31m symmetry in the setting as
described above and therefore the local coordinate system need not be modified.
OK to close the dialog window; OK; YES to save changes
This process is irreversible because atoms in the pattern molecule are deleted if they become
duplicit with respect to the introduced point group

30. Check atoms of the model tertiary butyl

Start “Edit Atoms”;
The number of independent atoms is now only 5.
Select the first atom of the model tertiary butyl and make double-click;
Switch to Edit to see the site occupancy;
You should get:
N31 0.166667
C81 0.166667
C91 0.500000
H1c91 0.500000
H2C91 1.000000

31. Refinement of the structure with local symmetry 31m – atomic positions of the tertiary
butyl fixed
Start the refinement;
Refinement converges with R(obs) value about 6.6% for 556 parameters, i.e. the same value
like before, because the shape of the pattern molecule is still fixed.

32. Refinement of the structure with local symmetry 31m

Right-click the icon of Refine; select tab “Various”;
[On the screen: Refinement commands]
Press the button “Fixed commands”;
Disable the fixed command; OK
Start the refinement;
Refinement converges with R(obs) value about 6.5% for 564 parameters.

328 Jana2006 Cookbook, version October 2017

Example 8.2: Dinitros
Application of rigid body groups for disordered parts of crystal structure
Revised: 3 March 2015

(3aRS,7aRS )-1,3-dinitrosooctahydro-1H-benzimidazole
Chemical formula: C7H12N4O2

Single crystal data measured with Oxford Diffraction four-circle diffractometer

Input files: dinitros.hkl, dinitros.cif_od
Frame scaling, absorption correction: done with software of the diffractometer

1. Creating new jobname

Start Jana2006
“File → Structure → New” opens a file manager
Left pane: locate directory with input files
Right pane: double-click dinitros

2. Import Wizard
Select “Known diffractometer formats”; NEXT
Select “Oxford Diffraction – CCD”; NEXT
Leave all settings unchanged; NEXT
Leave all settings unchanged; NEXT
For absorption correction select “None or done before importing”; NEXT
FINISH

3. Data Repository
Each line in data repository window corresponds to one data set.
OK; YES to accept the data set

4. Symmetry Wizard
[On the screen: “Tolerances for crystal system recognition”.]
Leave all settings default; NEXT
[On the screen: Select Laue symmetry]
Select Monoclinic; NEXT
Select primitive unit cell; NEXT
[On the screen: Select space group]
Select P2/c; NEXT

329 Jana2006 Cookbook, version October 2017

 Accept the space group; FINISH
[On the screen: define space group window]
OK to accept the space group P2/c

5. Creating refinement reflection filed

NEXT to confirm threshold 3sigma
[On the screen: settings for merging of reflections]
Leave all settings default; NEXT
Accept the data block and calculate coverage; FINISH

6. Structure Solution Wizard

[On the screen: window of Structure solution]
In “Formula” textbox type list of chemical elements for Dinitros: C7 H12 N4 O2;
Set "Formula units" to 4 and calculate density;
The density 1.393 g/cm3 looks reasonably so probably Z=4 is correct
Select “use Superflip”, “Peaks from Jana2006”; leave other settings default; “Run
solution”
Close listing of Superflip
Superflip converges with R value 16% and confirms symmetry P2/c
Press “Accept last solution” to leave the Structure solution wizard

7. Verification of the structure solution

Start “Plot structure”; “Draw + Continue”
[On the screen: window of Diamond]
Start “Build → Get molecules” or use on the bottom toolbar
Rotate the molecule ( and on the bottom toolbar) to get an optimal view
[On the screen: plot of the two independent molecules]

The results may be slightly different for each run of Superflip, because it starts from
random phases.
Draw rectangle to select all atoms of the molecule or press Ctrl-A
Right-click on one of the selected atoms, choose “Add → Atom labels”
[On the screen: dialogue of Diamond for atom labeling]
Note incorrect chemical types (using the chemical diagram above);
Quit Diamond;
Run “Edit atoms”;
Correct the improper chemical types;
Rename atoms according to chemical types;
Make symmetry contiguous motifs;
Quit “Edit atoms”, save the changes
For this structure making the symmetry contiguous motifs is very important. To make sure
the motifs have been correctly created:

330 Jana2006 Cookbook, version October 2017

 Return to Diamond;
Use or “Build → Add all atoms”;
Connect atoms with bonds
This will plot symmetry independent atoms of the structure model without any symmetry
transformation. They should create motifs:

If these motifs are not created the atoms must be put to the motif manually using
symmetry transformations in "Edit atoms". Without proper motifs this example will not
work!
Quit Diamond

8. Refinement
[On the screen: basic window of Jana]
Right-click the icon of Refine.
[On the screen: Refinement commands]
Define 100 of cycles; OK
Choose “YES+START”
Refinement converges with R(obs) value about 13%
Plot the structure
It should be the same like before
Start “Edit atoms” and change ADP of all atoms to “harmonic”
Repeat refinement
Refinement converges with R(obs) value about 9.4%
Plot the structure
Activate ADP ellipsoids using “Picture → Model and Radii”
One of the molecules has very large ellipsoids for carbon atoms:

9. Labeling
At this point we would need some more practical labeling of atoms. Jana2006 doesn’t
include tools for graphical renaming. The labels and their changes must be planned in
Diamond, noted and realized in “Edit atoms” tool or by direct editing of M40
The desired labeling for the molecule with large ADP’s (starts from C1) and the one with
smaller ADP’s (starts from C5):

331 Jana2006 Cookbook, version October 2017

 0.75C21C4‘
(The molecule with large ADP’s starts from C1)
Plot the molecules in Diamond; activate atom labels;
EITHER note the desired changes and Quit Diamond;
OR rename atoms in Jana2006 running parallel with Diamond (Diamond started with
“Draw+continue”)
(You may also keep Diamond window open and type the new labels directly to Edit atoms
tool of Jana2006)
Start “Edit atoms”; select all atoms and choose “Rename the selected atoms manually”
Plot the structure and verify the labels

10. Splitting positions with large ADP

The largest ADP have been found for atoms C2 and C4
Start “Edit atoms” and split positions of C2 and C4, using default options for splitting

The splitting procedure created pairs C2-C2’ and C4-C4’, their ADP are isotropic and the
same for each pair, partial occupations are ½ and the sum of occupancies for each pair is
kept by a Restriction for the case it would be refined
Run refinement
Refinement converges with R(obs) value about 10%
Start “Edit atoms” and activate refinement of occupancy for C2 and C4
Run refinement
Refinement converges with R(obs) value about 8.4%
The refined occupancies in both pairs are in the ratio 1:3, i.e. one of the split atom in each
pair is much weaker.
Plot the structure; distinguish by colors the weakly occupied carbon atoms
Activate plotting of ADP ellipsoids and plotting of labels

332 Jana2006 Cookbook, version October 2017

 In the plot C4, C2 have occupancy ~0.25; C4’, C2’ have occupancy ~0.75, the other atoms
have a full occupancy. We can distinguish two chairs: C4-C4-C3-C2-C2-C3 and C4’-C4’-C3-
C2’-C2’-C3, the latter having stronger occupation.

Warning: splitting of atomic positions was automatic process. In particular solution it may
happen that the labels of the strongly and weakly occupied atoms (i.e. atoms with and
without comma) will be exchanged. In such a case we recommend relabeling the atoms
according to this plot in order to follow easily the cookbook.

11. Rigid body description

Refinement of the weakly occupied positions is unreliable because ¼ of carbon is an
electron density comparable with hydrogen. In this part we will describe the disorder with
a rigid body. For model molecule we will take the stronger version of the molecule.

Start “File → Structure → Save As” and create a backup copy

Start “Parameters → Molecules → New molecule”
Choose the name for the molecule, e.g. “molec”
Select “Atomic part”; NEXT
[On the screen: Select atoms for the molecule]
Select O1, N1, N2, C1, C2’,C3,C4’
OK
The selected atoms will be removed from the atomic part of the structure model, and they
will be used as a pattern molecule. This pattern molecule will be later on transformed to
the actual places in the crystal structure by a translation vector and three rotations
[On the screen: Define the molecular reference point]
Reference point of the molecule is used for calculation of symmetry restrictions. In our
case C1 is located in a special position while the other atoms not. Therefore C1 should be
the reference point
Select “Explicit”; for “Reference point” type C1; NEXT
[On the screen: complete molecular position #1]
For the placement of the model molecule to the first actual position the program shows

333 Jana2006 Cookbook, version October 2017

zero rotations and zero translation. This is because the model molecule has been taken
from the existing atoms in the atomic part of the structure model. “Coincidence ratio 7/7”
means that 7 atoms of the model molecule placed to the actual position #1 coincides with
existing seven atoms in the atomic part. This is correct - the 7 atoms have been used for
the model molecule.

’
For “Occupancy” type 0.75;
“Apply + Next position”
[On the screen: Define and complete molecular position #2]
In this dialogue we will define rotations and a translation vector for placing the model
molecule to the second actual position. This is done by defining three atoms of the model
molecule and three atoms in the atomic part of the structure model, which should
coincide. In order to get correct placement of the second position the split atom positions
must be used.
Use C1, C2’, C4’ as model atoms and C1, C2 and C4 for the actual position
Press “Show coinciding atoms”

We expected that the 7 atoms in the actual position will coincide with 7 atoms in the
atomic part. However, we obtained only 3 coinciding atoms .
OK to quit information window
Select “Apply inversion”; Press “Show coinciding atoms”

334 Jana2006 Cookbook, version October 2017

Now we have seven coinciding atoms. Thus to get the second position the rotation must be
combined with inversion. The inversion centre is the reference point.
For “Occupancy” type 0.25
“Apply + End”
Start “Edit atoms”
The atoms of the model molecule are blue
Rename C2’ to C2 and C4’ to C4
Change occupancy of C2 and C4 to 1 (the model molecule has the full occupancy)
Change ADP of C2, C4 to “harmonic”
Quit “Edit atoms”
Plot the structure
Before starting Diamond Jana transforms the molecular positions back to the atomic part.
Atoms of the first position will obtain the suffix “a”, the second position has the suffix “b”

Note that all positions except C1 are now more or less split. Close positions in the rigid
body model are not problematic because the individual atomic parameters will be refined
only once for the model molecule.

Start “Parameters → Molecules → Edit”

Select refinement of “aimol” (this is occupancy of the molecular position)
OK; YES to save changes
Start Refinement options
In page “Various” go to “Restrictions” and delete two existing commands
The deleted restrictions concerned the split atomic positions and they are no more needed
In the same form define a new restriction command between Molec#1 and Molec#2:
coordinates and ADP parameters will not be restricted but molecular occupancies will
335 Jana2006 Cookbook, version October 2017
 keep the overall sum.
OK; OK; Yes+Start
Refinement converges with R(obs) value about 8%
The R value is still rather high but in a light structure missing hydrogen atoms may play an
important role

12. Local coordinate system

Atoms in the atomic part are generated according to the space group symmetry. In order
to generate atoms within a model molecule, we have to define local symmetry of the
molecule. After the placement of the model molecule to the final position in the unit cell,
the local symmetry and the space group symmetry are combined.

In our case, if we define the local symmetry 1, still keeping only one half of atoms in the
model molecule, the molecule would be completed after correct placement on a two-fold
crystallographic axis. However, we could not keep hydrogens of the carbon C1 in
geometrically expected positions because the keep command applies to the model
molecule and hydrogen positions need to be generated by symmetry. For this reason, we
have to define the local symmetry.

In our case, the model molecule has a two-fold axis along b and its point group will be 2.
Because the internal definition of the point group expects rotation along c we have to
define the point group 2 and change the local coordinate system.

Start “Parameters → Molecules → Edit”

For “point group” type “2”; TAB
Check “use local system” and define “1 0 0” for x and “0 1 0” for z

OK; OK; YES to save changes

Run refinement – the results should be the same like without the local symmetry

13. Hydrogen atoms

Start “Edit atoms”
Select all carbon atoms including the ones of the model molecule
Add hydrogen atoms in automatic mode
Verify by plotting placement of hydrogen atoms
For suppressing one of the disordered positions its molecular occumapcy can be
temporarily set to zero through "Parameters → Molecules → Edit". In such case, do not
forget to set it back after the plotting.

336 Jana2006 Cookbook, version October 2017

Run refinement
R value drops to 3.9%, GOF=2.6.

Start EditM50
Go to the page “Composition”
Press “Formula from M40”
It should print C7 H12 N4 O2 for Z=4
Start Refinement options, page Basic
Change Instability factor from 0.01 to 0.02
Run refinement
Refinement converges with R(obs)=3.9%, GOF=2.0.
Ideally, the instability factor should be determined for a given instrument. In the practice
we know that it ranges between 0.01 and 0.02.
GOF about 2.0 may indicate that disorder is not described exactly. It may also indicate that
the data contains information about bonding effects.
Start Fourier options
In page Basic define “Map type” “Fobs-Fcalc difference Fourier”
Run difference Fourier
In the listing of Fourier check list of maxima and minima
The residua are small like for fully solved structure
Last step, which is skipped in this example would be comparison of the molecular model
with a model built from split atom positions. The resulting R value would be very similar to
that of the molecular model.

337 Jana2006 Cookbook, version October 2017

338 Jana2006 Cookbook, version October 2017
Example 9.1: L9A
Getting familiar with the output of Superflip. Some useful tricks.

Revised: 15 September 2014

Single crystal data measured with Oxford Diffraction four-circle diffractometer

Input files: L9A.m40, L9A.m50, L9A.m90, L9A.m95

The purpose of this exercise is to understand the log-file of Superflip, point out the most
important information to check, and to learn some basic tricks for the case Superflip does not
converge with default parameters. The structure used for the exercise is an organic structure
with symmetry Pna21.

1. Load the structure

The structure is already prepared in Jana format for your convenience.

Start Jana2006;
“File → Structure → Open” opens a file manager;
Left pane: locate directory with input files;
Right pane: double-click L9A;

Plot the structure, if you wish, check EditM50 to see the composition and symmetry.

2. Run Superflip.

Double click the icon Structure solution

339 Jana2006 Cookbook, version October 2017

[On the screen: Structure solution]

Select “use Superflip”;

Start “Run Solution”;
Superflip starts, and after a few moments you will see the Superflip’s log-file open in the
window of Jana. It should be very similar to the commented file below. It will not be exactly
the same, because the iteration starts from random starting point, but the differences
should be minor. Please take your time to go through the text below, read the comments,
try to understand the listing both in this exercise and on your screen. Are the two listings
different in any important manner?

The first part is the header of the file, with information about the version. If you encounter
strange behavior of Superflip, make sure in http://superflip.fzu.cz you have the newest
version.The executable of Superflip downloaded from this web page should be copied to the
SUPERFLIP directory of Jana2006 installation.
**************************************************************
Superflip
Version: 12/04/13 7:57
Palatinus, L. & Chapuis, G.(2007): J. Appl. Cryst. 40, 786-790
http://superspace.epfl.ch/superflip
**************************************************************

Next part is a review of input parameters and crystallographic information. Superflip

currently writes only the matrix-vector representation of the space group, and not the space
group symbol.

-----------------------------------------------
Start of the calculation: 28.JAN 2014, 14:29:49
-----------------------------------------------

###################################################################
# Following data were read from the input file or set as default: #
###################################################################

Job title: L9A

------------------------

340 Jana2006 Cookbook, version October 2017

Information about files:
------------------------
Name of the input file: L9A.inflip
Density will be writen in following files:
L9A.m81, with type jana
L9A.m80, with type m80
Warning: If the outputfile exists, it will be overwritten by the new density.
Logfile will contain only basic information about the calculation.
Logfile will contain information about the data coverage.

-----------------------------
Crystallographic information:
-----------------------------
Superspace dimension: 3
Dimension of the physical space: 3
Direct cell parameters: 15.3079 9.2915 13.8059 90.0000 90.0000 90.0000 Volume: 1963.65942
Reciprocal cell parameters: 0.0653 0.1076 0.0724 90.0000 90.0000 90.0000 Volume: 0.00051
Wilson plot will be based on this composition: C84H104N8O16
4 symmetry operations found, their list follows:

1: 1 0 0 0.0000 2: -1 0 0 0.0000 3: 1 0 0 0.5000

0 1 0 0.0000 0 -1 0 0.0000 0 -1 0 0.5000
0 0 1 0.0000 0 0 1 0.5000 0 0 1 0.0000

4: -1 0 0 0.5000
0 1 0 0.5000
0 0 1 0.5000

The structure is non-centrosymmetric.

--------------------------
Settings of the algorithm:
--------------------------
Number of voxels: 40 24 36 Total: 34560
Number of voxels for polishing: 80 48 72 Total: 276480
The density modification method will be charge flipping.
Delta will be determined automatically.
The iteration will be stopped when the convergence is detected or after ***** cycles.
After the regular iteration 5 cycles of "density polishing" will be applied.
The random number generator will be initialized automatically.

Random initial phases will be assigned to the structure factors.

Isotropic Debye-Waller factor: 0.000
Proportion of reflections to be treated as weak: 0.000
All reflections with sin(th)/lambda < 0.400 not present in the input data set will be treated a missing.
The amplitudes of the missing reflections will be left unconstrained.

The resulting density will be shifted and averaged according to the symmetry operations given above.
Following symmetry operations will be used to locate the origin of symmetry: 3 4

Most of the information above is just diagnostic and is not important for ordinary user.
However, the number of reflections treated as weak might be changed, if Superflip does not
converge with default parameters. See below for more discussion on this parameter.

Next part is an overview of reflection import. “Number of reflections treated as missing” is

the number of low-resolution reflections that are not present in the input data. By default,
Superflip lets their value vary freely during the iteration. However, if this number is large
(say more than 10% of all reflections), the iteration may fail. In such a case a more
sophisticated treatment of reflections (or a more complete dataset) may be necessary. See
manual of Superflip, keyword missing for more information.

341 Jana2006 Cookbook, version October 2017

##################################
# Information about reflections: #
##################################

Number of reflections in the input file: 1624

-------------------------
Averaging of reflections:
-------------------------
Redundancy : 1.000
Rint : -----

Maximum indices in expanded reflection set: 17 10 15

Number of reflections treated as missing: 6

Wilson plot is always a useful indicator of the data quality and consistency. If you encounter
problems with your structure, do not forget to check the Wilson plot and make sure it does
not show anomalies. The value Biso is the Wilson-plot based estimate of the average Debye-
Waller factor. If it is larger than 2, a correction of the input data might be needed. We will
learn how to make this correction later in this exercise.
------------
Wilson Plot:
------------

circles: <I>/SUM(f^2)
dots : linear fit y = 1.5450 + -3.480 * [sin(theta)/lambda]^2

4.15377-| o o
-| . . o
Scale^2 -| o o . o o
-| o o . o
-| . o o o
2.99073-| o . . o o
-| o . . o
-| . . o
-| o . .
-| o o . .
2.15334-| . . o
-| . o
-| o o
-| .
-| o . .
1.55041-| . .
-| o . . o
-| o o .
-| o
-| o
1.11630-| o
I-------------------I-------------------I-------------------I----------
0.00 0.10 0.20 0.30
[sin(theta)/lambda]^2

Scale (F_obs = K * F_abs): 2.165

Biso: 1.740

Coverage statistics shows the coverage as a function of resolution. The coverage should be
high (optimally more than 90%) up to a reasonable resolution (dmin<1.1 or better). If one of
these conditions is not fulfilled, Superflip may fail. Very low coverage at low resolution is
especially dangerous.
----------------------------------------------------------
Coverage statistics of the expanded reflections by shells:
----------------------------------------------------------
Resolution (sin(th)/l): 0.050 0.100 0.150 0.200 0.250 0.300 0.350 0.400
Resolution (d_min): 10.000 5.000 3.333 2.500 2.000 1.667 1.429 1.250
Obs. refl. in shell: 0 25 69 139 238 349 515 655
Total refl. in shell: 0 27 73 139 238 349 515 655
Coverage in shell: 0.0% 92.6% 94.5% 100.0% 100.0% 100.0% 100.0% 100.0%
Commulative coverage: 0.0% 92.6% 94.0% 97.5% 98.7% 99.3% 99.6% 99.7%

Resolution (sin(th)/l): 0.450 0.500 0.550 0.575

Resolution (d_min): 1.111 1.000 0.909 0.870
Obs. refl. in shell: 876 1093 1323 708
Total refl. in shell: 876 1093 1323 730
Coverage in shell: 100.0% 100.0% 100.0% 97.0%
Commulative coverage: 99.8% 99.8% 99.9% 99.5%

342 Jana2006 Cookbook, version October 2017

 Iteration record shows some most important quantities during the iteration process. At the
beginning the program tries to estimate the value of delta automatically. This procedure is
very reliable, and only in very rare cases a manual selection of delta is necessary. The four
quantities shown in every line are as follows:
- R: The R-value between the observed structure-factor amplitudes, and the amplitudes
calculated after flipping. This value usually lowers at the convergence.
- Charge: Total charge in the current density. This value progressively lowers during the whole
iteration, but drops down abruptly at the convergence
- Peaks: The third central moment of the density – a measure of “peakiness” of the density.
This value grows at the convergence.
- Score: A combined “Figure of merit” calculated from the three quantities above as well as
from their derivatives. If the score grows above 0.8, Superflip considers the iteration as
converged.
Note: An automatic detection of convergence is an extremely difficult task, because it must
cover a wide range of cases. Therefore, despite of a substantial effort invested into the
detection, Superflip occasionally fails to detect convergence. It is therefore always worth
checking whether the structure is solved, even if the iteration does not seem to converge.
#############
# Iteration #
#############

Estimated delta: 2.2715

Random seed: 131727704

Searching for a proper delta:

Current delta = 2.27153 (k_ed = 1.00000)
Total/flipped ratio = 1.021. Increasing delta.

Current delta = 2.43053 (k_ed = 1.07000)

Total/flipped ratio = 0.856.
Criterion for delta fulfilled, continuing iteration.
20 R: 49.925 Charge: 1220.10 Peaks: 1.26 Score: ---
30 R: 48.103 Charge: 1149.05 Peaks: 1.36 Score: ---
40 R: 49.758 Charge: 1128.06 Peaks: 1.47 Score: ---
50 R: 49.224 Charge: 1120.09 Peaks: 1.53 Score: ---
60 R: 49.160 Charge: 1067.55 Peaks: 1.58 Score: ---
70 R: 49.279 Charge: 1092.28 Peaks: 1.62 Score: ---
80 R: 48.642 Charge: 1088.73 Peaks: 1.70 Score: ---
90 R: 48.457 Charge: 1063.52 Peaks: 1.74 Score: ---
100 R: 48.313 Charge: 1065.67 Peaks: 1.74 Score: ---
200 R: 46.931 Charge: 1057.27 Peaks: 1.89 Score: ---
300 R: 38.686 Charge: 972.57 Peaks: 2.09 Score: 0.15

Calculation successfully converged after 390 cycles.

Last iteration record:
390 R: 23.891 Charge: 878.59 Peaks: 2.23 Score: 0.91

5 cycles of noise suppression follow:

5 R: 14.793 Charge: 0.00 Peaks: 2.71

At this moment the calculation is converged (or the maximum number of iteration cycles
was reached). Superflip has the capacity to check the resulting density for symmetry. It is
advisable to always check this part of the log-file, and to make sure that the detected
symmetry corresponds to your expectations. Derived space group P1 in cases where higher
symmetry is expected usually indicates that the structure is not solved.
Superflip lists all symmetry operations compatible with the lattice. Each line in the listing
corresponds to one symmetry operation. The symbol of the operation is followed by a one-
line representation as used in International Tables for crystallography. Next is the
“symmetry agreement factor”. This number indicates quantitatively, how well the operation
is present in the electron density. Numbers below 10 indicate a very good match, number
between 10 and 25 are still probable symmetry operations of the density, and numbers
above 25 usually mean that the corresponding symmetry operation is not present. In the
343 Jana2006 Cookbook, version October 2017
example below the n-glide, a-glide, and two-fold screw axis have agreement factors below
0.1, indicating an excellent agreement. All the rest of the list has a far too high value. The
rows of XXXXXX after the symmetry agreement factor are just a graphical representation of
the agreement factor, and serve as a quick guide for eye.
#####################################
# Checking the density for symmetry #
#####################################

Symmetry operations compatible with the lattice and centering:

Symmetry operation agreement factor
n(1,0,0): -x1 1/2+x2 1/2+x3 0.022 XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
a(0,1,0): 1/2+x1 -x2 x3 0.072 XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
2_1(0,0,1): -x1 -x2 1/2+x3 0.090 XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
m(0,0,1): x1 x2 -x3 74.885 XXXXXXXXXXXXX
2_1(0,1,0): -x1 1/2+x2 -x3 74.894 XXXXXXXXXXXXX
2_1(1,0,0): 1/2+x1 -x2 -x3 74.907 XXXXXXXXXXXXX
m(1,0,0): -x1 x2 x3 78.690 XXXXXXXXXXX
2(0,0,1): -x1 -x2 x3 78.739 XXXXXXXXXXX
2(0,1,0): -x1 x2 -x3 80.683 XXXXXXXXXX
-1: -x1 -x2 -x3 80.741 XXXXXXXXXX
b(0,0,1): x1 1/2+x2 -x3 84.318 XXXXXXXX
n(0,0,1): 1/2+x1 1/2+x2 -x3 91.302 XXXX
2(1,0,0): x1 -x2 -x3 91.311 XXXX
a(0,0,1): 1/2+x1 x2 -x3 91.487 XXXX
b(1,0,0): -x1 1/2+x2 x3 92.258 XXXX
c(0,1,0): x1 -x2 1/2+x3 104.952 X
m(0,1,0): x1 -x2 x3 107.481 X
c(1,0,0): -x1 x2 1/2+x3 120.115 X
n(0,1,0): 1/2+x1 -x2 1/2+x3 135.532 X

An experienced crystallographer can “see” the space group already in the list of the
symmetry operations, but Superflip can analyze the list automatically. The default threshold
for accepting a symmetry operation is the symmetry agreement factor below 25. All such
operations are combined together, and a final space group is printed as Herman-Mauguin
symbol, Hall symbol, Fingerprint, and a list of symmetry operations. The “Fingerprint” is an
internal space-group symbol used by Superflip for a database search, and currently does not
have any other use.
-------------------------------------------------
Space group derived from the symmetry operations:
-------------------------------------------------
HM symbol: Pna21
Hall symbol: P 2c -2n
Fingerprint: 3300223}240qY3 (3/4,3/4,0)
Symmetry operations:
1: x1 x2 x3
2_1(0,0,1): -x1 -x2 1/2+x3
n(1,0,0): 1/2-x1 1/2+x2 1/2+x3
a(0,1,0): 1/2+x1 1/2-x2 x3

As the density is reconstructed in P1, the origin is arbitrary. To allow for further processing,
the density must be shifted to the standard origin of the space group, and averaged to have
the desired symmetry exactly. The lines below show the record of this process.

Important note: The “Check for symmetry” described above is just indicative, and does not
have any impact on further symmetry processing (unless you force it explicitly in the input
file of Superflip – see manual, keyword derivesymmetry for details). Thus, the output
symmetry will be the symmetry indicated in the input file. If you call Superflip from Jana with
default parameters, this is the symmetry recorded in m50.

“Agreement factors of individual generators” show agreement factors for the generators of
the input space groups. These generators are used to locate the origin of the space group. If
the input space group is incorrect, or the solution did not converge, these agreement factors
will be high (more than 25). Low values of the agreement factors are another reliable
indicator that the structure was correctly solved.

344 Jana2006 Cookbook, version October 2017

 ############################################
# Search for the origin of the space group #
############################################

Position of the origin in the CF map:

0.6641 0.6559 0.0000
Agreement factors of individual generators:
Number smb agreement
3 a 0.07
4 n 0.02

Overall agreement factor: 0.06

These are the final notes in the log file with information about the file with the output:
You can obtain more information about the reconstructed reflection phases by using 'expandedlog yes'.

Electron density written to file L9A.m81.

---------------------------------------------
End of the calculation: 18.JAN 2010, 13:17:29
---------------------------------------------
Superflip version: 07/16/09 09:12

Press “Close” to close the listing and return to the Structure solution dialogue

3. View the output of Superflip and resulting structure model

[On the screen: Structure solution]

After closing the listing of Superflip, Jana2006 automatically interprets the resulting density
in terms of atomic positions and types. For chemical types, only the ones available in
“Formula” are considered.
Press “Draw structure” to visualize structure model interpreted by Jana2006 from the
density provided by Superflip
Note that the structure model has the symmetry originally det by the user. If the symmetry
suggested by Superflip is different and you want to use it, it must be set by the “Change the
space group” button, and the interpretation of the density map or complete charge flipping
must be repeated.
“Return to previous” loads the previous structure model. Note that there is only one step
back possible.

If everything goes well, the structure model correctly describes the density produced by
Superflip. However, in some cases this can lead you to the wrong impression that this is
what Superflip produces. In case of problems (typically disorder or wrong chemical types in
the formula) viewing the density may be very helpful.

Press “Draw 3d map” to view the 3D isosurface representation of the raw density produced
by Superflip
The crystallographic (not only) viewer VESTA must be installed on your computer and
enabled it in Jana. If not, consult the installation notes and install the program first. Jana
passes the volumetric data to VESTA along with the structure model. VESTA shows an
overlay of both, with the density represented by isosurfaces, and atoms by spheres. In this
case the density shows nice spherical maxima at the positions of atoms. All maxima are
“occupied” by an atom. Thus, the structure model is likely to be complete and ready to be
refined.

345 Jana2006 Cookbook, version October 2017

 By default, the density from Superflip is saved in the file jobname.m81. That means that you
can directly view it even after quitting the Structure solution dialogue as long as you do not
overwrite it by calculating another Fourier map. M81 can be viewed with Contour program.

By default, the structure factors from Superflip are saved in the file jobname.m80. That
means the Fourier map can be recalculated by Jana2006 as long as m80 is not overwritten
by refinement.

4. Modify the default settings for Superflip

The default settings passed to Superflip by Jana are applicable to most standard cases.
However, under some circumstances the change is necessary. The most useful changes will
be listed here.

Hint 1: use normalized intensities:

For more difficult structures using normalized intensities helps enormously.
Check the checkbox next to Use local normalization:

Hint 2: Run Superflip several times with a limited number of cycles:

Check the checkbox “Repeat Superflip: number of runs”

You may select a different number of repetitions than the default 10.
If necessary, modify the maximum number of cycles per run, for example to 5000:

This option may help if the solution appears unstable or if it gets trapped in a local
minimum. Some of the repeated runs may escape from the local minimum. Superflip selects
the best of all attempts for the output.

Hint 3: Use a non-random starting model

Select the option “Patterson superposition map” from the list named “Starting model”:

With this option Superflip first calculates a Patterson superposition map and then uses this
map as a starting model for the iteration. Very often this option speeds up the convergence,
and sometimes it makes the difference between solved and non-solved structure. However,
sometimes – especially for simpler structures – the convergence is so rapid that it cannot be
detected and Superflip keeps running. This option saves time only for larger structures,
ideally with at leas one non-light atom.

Hint 4: use the AAR algorithm:

Select the AAR iteration scheme from the list of options at the bottom of the window:

The AAR algorithm appears to perform better than the standard algorithm for very
incomplete data or for low-resolution data. However, it is less stable than the standard
algorithm and it is more difficult to detect its convergence. AAR is the default algorithm for
electron diffraction data.
346 Jana2006 Cookbook, version October 2017
Try one or more of the above hints and see how they affect the speed of convergence. You
should come to the conclusion that for this simple structure there is hardly any difference
between the options, but most of them speed up the convergence., However, combination
AAR + local normalization is too aggressive and does not lead to the solution.

Superflip offers a large number of other options, which are described in detail in the manual.
If you want to use any of these advanced options, then:
Check the checkbox “Allow manual editing of the command file before start:

Click Run solution

Before Superflip starts, the input file opens in a text editor and you may edit the file and
include whatever additional options you wish.

The price to pay for faster convergence is the increased noise in the result, and sometimes
more difficult detection of convergence. Because the default parameters work usually quite
well and yield the best densities, we recommend using them for first attempts. Only if the
iteration does not converge, try modifying the parameters according to the hints described
above.

Press “Quit” to close the Structure solution tool

Note that there is no “escape” option. The original M40 which might be present before
starting this tool is being overwritten after each “Run solution” process and the only way
how to retain it is to use in proper time the “Return to previous” button.

347 Jana2006 Cookbook, version October 2017

5. Integration of Superflip to Jana2006
The following scheme shows how Superflip communicates with the rest of Jana2006. M95
contains raw data; M90 is refinement reflection file (merged reflections without extinct
reflections); M40 is structure model; M50 is basic crystal information; M80 are structure
factors from refinement or charge flipping; M81 is density map

Data Import
Output: M95

Symmetry wizard
Updates symmetry in M50

Create refinement reflection file

Output: M90

Structure solution
(Superflip calculates density, Jana interprets it)

Run Solution Change Change

Reads M90*) formula symmetry
Writes M80, M81 Updates
and M40 M50

Draw structure
Reads M50,M40

Draw 3d map
Reads M81,M40

Refinement Fourier Calculation Contour

Reads M90,M40,M50 Reads M80,M50 Reads
Writes M40,M80 Writes M81 M40,M50,M81

*) Superflip expands M90 to P1 symmetry. However, its results are to some extent
influenced by the symmetry set by the user with the Symmetry wizard, because M90
contains merged reflection and systematically extinct reflections are omitted

348 Jana2006 Cookbook, version October 2017

Example 9.2: pArNO2
Solving the centro-noncentro ambiguity with Superflip

This relatively simple structure illustrates the advantages of using Superflip to determine the
space group.

Revised: 16 September 2014

Single crystal data measured with Oxford Diffraction four-circle diffractometer

Input files: pArNO2.hkl, pArNO2.cif_od

1. Create new structure

Start Jana2006
“File → Structure → New” opens a file manager
Left pane: locate directory with input files
Right pane: double-click pArNO2

2. Import Wizard
Select “Known diffractometer formats”; NEXT
Select “Oxford Diffraction – CCD”; NEXT
Leave all settings unchanged; NEXT
OK to confirm the name of the basic input file
Leave all settings unchanged (note the cell parameters and thus the probable crystal
system); NEXT
Finish the import wizard with all settings default

3. Symmetry Wizard
[On screen: Information about the symmetry wizzard.]
NEXT to start the wizard;
Leave all settings default; NEXT; OK;
[On screen: Select Laue symmetry]
Jana2006 suggests orthorhombic symmetry based on Rint. In this case the choice appears
clear.
Keep the suggested selection; NEXT
Select primitive unit cell; NEXT
[On screen: Select space group]
You are now offered a list of candidate space groups sorted by their likelihood. Note that
the first two space groups Pnma and Pn21a differ only by the presence of inversion center
and they have the same FOM:

349 Jana2006 Cookbook, version October 2017

 Jana by default does not use the distribution of E-values to estimate, if the structure is
centrosymmetric or not. The main reason is that the same information can be obtained
from Superflip with higher reliability. At this point the space group choice can be made
either randomly or based on chemical knowledge. The choice of one of the two space
groups has no impact on the structure solution by Superflip.

Select Pnma; NEXT

Accept the space group; FINISH

4. Creating refinement reflection file

Proceed with the creation of refinement reflection file, leaving all settings default until
you see the following screen:

5. Consulting the intensity statistics

At this point we will interrupt the standard flow of the solution process to check the E-
statistics and estimate the presence or absence of the inversion center.
OK; Quit to escape the structure solution dialog.
[On screen: Main window]
First we have to define the chemical composition.The crystallized molecule has sum
formula C42 H43 N1 O6 S4 (see the end of the exercise for the picture of the molecule),
number of molecules per unit cell: 8. The solvent was acetonitrile – we do not know the
number of solvent molecules per formula unit, but we may ignore the solvent contribution
for now.

Double click the icon EditM50

Select the page “Composition”
Enter “C42 H43 N O6 S4” in the first field;

350 Jana2006 Cookbook, version October 2017

 Keep 8 for the number of formula units; OK; Yes;
[On screen: Main window]
Go to Run → Intensity statistics
[On screen: Wilson plot]
Check the Wilson plot. Click anywhere in the plot to continue;
Yes to open the listing;
[On screen: listing of intensity statistics]
This listing contains a large number of statistics related to intensity normalization. We will
focus only on the following table (found around the middle of the listing):

Experimental Theoretical

Value All data hkl 0kl h0l hk0 Acentric Centric Hypercen.

mod(E) 0.801 0.804 0.779 0.499 0.867 0.886 0.798 0.718

E**2 1.000 1.003 1.086 0.473 1.095 1.000 1.000 1.000
E**3 1.690 1.694 2.131 0.630 1.773 1.329 1.596 1.916
E**4 3.556 3.570 5.127 0.997 3.433 2.000 3.000 4.500
E**5 8.781 8.832 13.989 1.728 7.628 3.323 6.383 12.260
E**6 24.409 24.611 41.363 3.164 18.939 6.000 15.000 37.500
mod(E**2-1) 0.983 0.981 1.177 0.905 0.985 0.736 0.968 1.145
(E**2-1)**2 2.556 2.565 3.955 1.051 2.244 1.000 2.000 3.500
(E**2-1)**3 15.740 15.908 28.239 0.592 10.922 2.000 8.000 26.000

This table shows the numbers typically used to estimate the centrosymmetry. If you
compare the column All data with the theoretical values, you can see that all numbers
indicate a centric or hypercentric distribution. Our guess Pnma thus appears to be strongly
supported by the statistics
Click the button “Close” to close the listing and return to the main window.

6. Structure Solution

Double click the icon Structure solution

[On screen: window of Structure solution]
Leave all settings default;
Click the button “Run solution” to start Superflip
Superflip converges smoothly and the listing of the solution procedure opens in Jana.
[On screen: listing of Superflip]
Locate the symmetry analysis in the listing. It should look very similar to this (rows of
XXXX are omitted here to save space):

#####################################
# Checking the density for symmetry #
#####################################

Symmetry operations compatible with the lattice and centering:

Symmetry operation agreement factor
a(0,0,1): 1/2+x1 x2 -x3 0.229
n(1,0,0): -x1 1/2+x2 1/2+x3 0.651
2_1(0,1,0): -x1 1/2+x2 -x3 0.714
2_1(0,0,1): -x1 -x2 1/2+x3 51.947
-1: -x1 -x2 -x3 52.037
2(0,0,1): -x1 -x2 x3 66.366
351 Jana2006 Cookbook, version October 2017
 m(0,1,0): x1 -x2 x3 71.723
2_1(1,0,0): 1/2+x1 -x2 -x3 72.620
c(0,1,0): x1 -x2 1/2+x3 72.643
2(0,1,0): -x1 x2 -x3 78.403
m(1,0,0): -x1 x2 x3 78.485
2(1,0,0): x1 -x2 -x3 78.709
n(0,1,0): 1/2+x1 -x2 1/2+x3 78.768
m(0,0,1): x1 x2 -x3 82.199
b(1,0,0): -x1 1/2+x2 x3 82.308
a(0,1,0): 1/2+x1 -x2 x3 83.217
c(1,0,0): -x1 x2 1/2+x3 89.079
n(0,0,1): 1/2+x1 1/2+x2 -x3 126.255
b(0,0,1): x1 1/2+x2 -x3 131.166

-------------------------------------------------
Space group derived from the symmetry operations:
-------------------------------------------------
HM symbol: Pn21a
Hall symbol: P -2ac -2n
Fingerprint: 33002241Y40qY3 (3/4,0,3/4)
Symmetry operations:
1: x1 x2 x3
a(0,0,1): 1/2+x1 x2 1/2-x3
n(1,0,0): 1/2-x1 1/2+x2 1/2+x3
2_1(0,1,0): -x1 1/2+x2 -x3

Look at the table. You can see that the first three operators belonging to the Pn21a space
group have very low agreement factors (less than 1%) indicating a very good match. All
other operators have agreement factors above 50%, which is very poor. In particular the
mirror plane perpendicular to b has agreement factor 71.7%. It is thus clearly absent in the
structure. The space group is without doubt Pn21m.

6. Insert the correct space group and solve the structure

Although the solution by Superflip is correct, the space group in Jana is still wrong, and the
solution is thus not usable. We have to set the correct space group and rerun Superflip.
Close the view of the log file;
Click “Change the space group” in the “Structure solution” form;
Type “Pn21a” in the upper left box; OK; YES;
As the symmetry changed, the refinement reflection file must be recreated. Create the new
file with default settings.
Press “Run solution” to start Superflip again with the correct symmetry;
Superflip starts, and should again converge smoothly to a solution.
Close the log-file viewer;
Click “Draw 3D map”;
View the resulting electron density and the interpreted model. The structure is correctly
solved, but quite a few peaks have not been assigned an atom. We know the expected
composition and we may thus use the alternative peak-interpretation option – use of an
external program EDMA:
Close VESTA;
Select the radiobutton “EDMA – fixed composition”:

352 Jana2006 Cookbook, version October 2017

Check the checkbox “Use old solution and reinterpret it:

Click “Run solution” to interpret the previous solution with the new setting.
A new screen appers with the run-time information from EDMA: When EDMA finishes, a
listing of Superflip opens.
Close the listing and return to the structure solution dialog
Double click “Draw 3D map”
This time you sould see all maxima in the map occupied by an atom. Most atoms should
have a correctly assigned atomic type, although carbon, nitrogen and oxygen may be
occasionally swapped.
Here is the correct molecule for reference (hydrogens omitted for clarity):

353 Jana2006 Cookbook, version October 2017

354 Jana2006 Cookbook, version October 2017
 Example 9.3: flo19
A challenging case of difficult symmetry and difficult convergence detection

Revised: 18 February 2014

Input files: flo19.cif, flo19.hkl. The structure is an organic molecule built by four aromatic
rings bridged by one or two nitrogen atoms.

1. Load the structure from CIF

Start Jana2006;
“File → Structure → New”;
Locate the input files, and double click the name flo19 in the right pane;
Select the radiobutton “Structure: from CIF”; NEXT; OK to confirm the file flo19.cif
OK to the warning that the CIF does not contain symmetry information;
At this moment you loaded the basic crystallographic information. You can check the
composition and unit cell dimensions by double clicking the EditM50 icon and consulting the
card Cell and Composition.

2. Import the reflection file

“File → Reflection file → Import/modify the reflection file”
Select “Single crystal: reflection file corrected for LP and absorption”; NEXT;
Check that the filename is flo19.hkl and that the file type is SHELX on I; NEXT;
2x NEXT for the two following screens; OK; Finish; OK; YES to finish the import wizard

3. Estimate the symmetry

When the reflection import finishes, you are automatically offered the symmetry
determination form 1.
NEXT to start the wizard;
NEXT to confirm the default settings of the test;
OK to the information about no higher metric symmetry;
On the screen with Rint for different Laue groups make sure you understand Jana’s choice of
the mmm Laue group.
NEXT;
On the screen with centering statistics make again sure that you understand the meaning of
the numbers and the choice Jana2006 made.
NEXT;
At this moment you are presented with the list of possible space groups. The upper part of
the list looks like this:

1. 1If you accidentally close the wizard or if you want to repeat the test, you can access the
symmetry determination procedure through the EditM50 icon, card Symmetry.
355 Jana2006 Cookbook, version October 2017
 Select the space group P21221 and click “Details”;
In the listing you can find basic information about the space group and the list of observed
reflections that violate the rules for systematic absences. Check the details for several space
groups. Which space groups would you consider possible and which would be the most
likely?

Because the solution in Superflip is not sensitive to the choice of the space group symmetry,
we will choose just one of the space groups without any systematic absences to keep the
solution unbiased.

Select the space group Pmmm; NEXT; FINISH; OK; YES to finish the space group selection
We selected the space group Pmmm not because it is the most likely choice, but because it
does not induce any systematic absences. We will now process the reflections, and ask
Superflip to estimate the space group. If we selected a space group with systematic
absences, some reflections would be excluded from the reflection file, and the result from
Superflip could be biased.

4. Create refinement reflection file

When the space group determination finishes, you are automatically offered the “Create
refinement reflection file” tool. If you accidentally exit it, you can always return to it through
“File → Reflection file → Create refinement reflection file”
NEXT; OK; NEXT; OK; FINISH to create the refinement reflection file with all settings default

7. Run Superflip.
Next Jana offers you the dialog for structure solution. If you want to open this dialog later,
just double click the icon “Structure solution” in the main window of Jana.
Make sure Superflip is selected;
Leave other settings default; Click Run solution;
Superflip starts and runs for many thousand cycles. No convergence is detected.
Press “Interrupt” to stop Superflip;
The log file of Superflip opens. Check the symmetry determination part. You will see that no
symmetry operation has agreement factor lower than 25% and thus the derived space group
is P1. Because we expect higher symmery (although we do not know the exact space group),
it is a sign that the calculation did not converge.
356 Jana2006 Cookbook, version October 2017
 Inspection of the log file does not reveal any serious problem, but a series of hints shows
that the structure can be difficult to solve. The potential problems are that the coverage
starts dropping already at dmin≈1.1 Å, and that the coverage at low resolution is very low
(only 21% at dmin≈ 5 Å). It can be assumed that quite a few strong reflections are missing in
the dataset. The first change to try is to use normalized intensities. When using normalized
intensities, Superflip also automatically treats the missing reflections in a special way that
improves the performance in case of data with missing low-order reflections.
Close the view of the log file;
Check the checkbox “local normalization”
Click “Run solution”

Superflip starts and probably runs again for many thousand cycles without detecting
convergence 2.
If you look at the evolution of the convergence indictors, you will most probably see behavior
similar to this2,3:
20 R: 56.175 Charge: 20.84 Peaks: 1.11 Score: ---
30 R: 57.537 Charge: 20.13 Peaks: 1.16 Score: ---
40 R: 58.291 Charge: 19.03 Peaks: 1.15 Score: ---
50 R: 58.075 Charge: 19.08 Peaks: 1.22 Score: ---
60 R: 57.460 Charge: 18.65 Peaks: 1.22 Score: ---
70 R: 59.161 Charge: 18.79 Peaks: 1.32 Score: ---
80 R: 57.279 Charge: 18.88 Peaks: 1.33 Score: ---
90 R: 59.624 Charge: 18.89 Peaks: 1.39 Score: ---
100 R: 56.954 Charge: 18.29 Peaks: 1.41 Score: ---
200 R: 55.838 Charge: 18.37 Peaks: 1.29 Score: 0.15
300 R: 57.292 Charge: 18.12 Peaks: 1.35 Score: 0.15
400 R: 55.889 Charge: 17.89 Peaks: 1.32 Score: 0.15
500 R: 55.456 Charge: 18.50 Peaks: 1.32 Score: 0.15
600 R: 55.902 Charge: 18.88 Peaks: 1.22 Score: 0.15
700 R: 53.192 Charge: 18.50 Peaks: 1.24 Score: 0.15
800 R: 53.102 Charge: 18.23 Peaks: 1.27 Score: 0.15
900 R: 50.902 Charge: 17.41 Peaks: 1.23 Score: 0.15
1000 R: 45.968 Charge: 18.16 Peaks: 1.19 Score: 0.15
2000 R: 43.139 Charge: 18.59 Peaks: 1.12 Score: 0.15
3000 R: 42.466 Charge: 18.12 Peaks: 1.11 Score: 0.15
4000 R: 43.158 Charge: 19.33 Peaks: 1.16 Score: 0.15
5000 R: 43.369 Charge: 18.80 Peaks: 1.11 Score: 0.15
6000 R: 44.689 Charge: 18.11 Peaks: 1.12 Score: 0.15

Note the bold part. We can see a clear drop in R-value accompanied by no change in the
charge and a small decrease (instead of expected increase) of the Peaks indicator. Although
the change in R-value is clearly visible, together with the decrease in Peaks it is not enough
to trigger automatic convergence detection. However, it is likely that the iteration
converged.
Press “Interrupt” to stop Superflip;
In the log file locate the symmetry determination part.
You will see a listing similar to this:

2
Occasionally it may detect convergence, but in most cases it does not.
3
Rarely the calculation does not converge even after several thousand cycles. In that case interrupt Superflip and
run it again.
357 Jana2006 Cookbook, version October 2017
 Go carefully through the list, and make sure you understand, why Superflip suggests the
space group P21221. Note also that the space group has nothing to do with the Pmmm space
group used as input. This confirms that the symmetry derivation by Superflip is indeed
independent from the input space group.

Note also that the space group found by Superflip is the same as the most likely space group
suggested by Jana2006 on the basis of systematic absences. Thus, in this case the two
indications agree. However, the analysis of systematic absences was not conclusive, while
the result of Superflip leaves no doubt about the correct space group.

8. Insert the correct space group and solve the structure

Although the solution by Superflip is correct, the space group in Jana is still wrong, and the
solution is thus not usable. We have to set the correct space group and rerun Superflip.
Close the view of the log file;
Click “Change the space group” in the “Structure solution” form;
Type “P21221” in the upper left box; OK; YES;
As the symmetry changed, the refinement reflection file must be recreated. You will be
guided through the process. Repeat it as you did in step 4 of this exercise.

358 Jana2006 Cookbook, version October 2017

Press “Run solution” to start Superflip again with the correct symmetry;
Superflip starts, and should again converge smoothly to a solution. You may again need to
interrupt the iteration, when you see the drop of the R-value. Check that the derived
symmetry is again P21221.
Close the log-file viewer;
Double click “Plot structure”; Draw+continue
In Diamond click the “Get molecules directly” button in the bottom toolbar. The
molecule should look like this, possibly with a few carbon and nitrogen atom types
interchanged:

Close Diamond;
Double click the icon “Refine”
The refinement will converge to an R-value around 10% or less depending on the number of
incorrectly identified atoms. The structure is solved. Now the atoms must be correctly
assigned to chemical types, and the structure refined in the standard manner.

359 Jana2006 Cookbook, version October 2017

360 Jana2006 Cookbook, version October 2017
Example 9.4: MnVPO
A difficult case for symmetry determination

Revised: 12 February 2014

(Mn(H2O)4)(VO)2(PO4)2
Single crystal data measured with Oxford Diffraction four-circle diffractometer
Input files: MnVPO.m50, MnVPO.m95

1. Opening the structure

Start Jana2006
“File → Structure → Open” opens a file manager
Left pane: locate directory with input files
Right pane: double-click MnVPO_start

You opened a structure which contains crystal information about the compound and
hkl-data. There is no structural information present yet. To preserve the original files, save
the structure under another name:

“File → Structure → Save as” opens a file manager

Type “MnVPO” in the line under the right pane and press Enter
“YES” to the question if you want to continue with the new structure

2. Symmetry Wizard
“File → Reflection file → Make space group test”
Proceed with the space group test. Check the table of the Laue groups, and note Rint for Laue
groups 4/m and 4/mmm. Continue until the screen with the title “Select space group”. You
should see this output:

In this case it is impossible to determine the space group only from systematic absences.
Select the space group I4/m, NEXT, FINISH

3. Creating refinement reflection file

Answer YES to the question if you want to create refinement reflection file
Leave all settings default; NEXT; OK; NEXT; OK;
Accept the data block; FINISH

361 Jana2006 Cookbook, version October 2017

4. Structure Solution
Click the icon “Structure solution” in the main window of Jana

Make sure that the formula is H8 Mn O14 P2 V2 with Z=2

Leave other settings default; Click the button “Run solution”
[On the screen: listing of Superflip]
Superflip converges rapidly. The symmetry determination part of the log-file looks like this:

Superflip suggests the space group I4/mmm. This space group is incompatible with the
detected Laue group (Check your notes with Rint of the Laue groups 4/m and 4/mmm). It is
a supergroup of all three possible space groups suggested by the symmetry wizard. This is
an indication that the structure is peculiar in some respect.

Close listing of Superflip

Quit the Structure solution dialog

Note: Superflip suggests a different space group, but it does not use this space group to
produce the final density. The final result has always the symmetry defined in EditM50. If
you want to accept the symmetry suggested by Superflip, you have to change the symmetry
and rerun Superflip to get new solution.

362 Jana2006 Cookbook, version October 2017

 “Parameters → Atoms → Edit”
Check that the list of atoms contains one Mn, V and P atom and three oxygens. The peak
search sometimes does not assign all atoms at once. In practice the missing atom would be
located among the Fourier peaks. For the purposes of this example, however, make sure the
solution contains 6 atoms immediately to keep your files compatible with the rest of the
exercise. If it does not, rerun Superflip again.

5. View the structure

Double click the icon “Plot structure”; Answer NO to question about Fourier peaks;
In Diamond click the button for filling one unit cell and for filling coordination spheres
;
You should get a view similar to this:

The structure looks correct. To convince yourself that the cations have correct coordinations,
rotate the structure and increase the accepted V-O bond length to 2 (“Build →
Connectivity”). Try to understand, why Superflip suggested a wrong space group before you
go to the next page and read the solution.

363 Jana2006 Cookbook, version October 2017

 The reason Superlip gets confused by this structure is the strong pseudosymmetry. All atoms
except for one oxygen (the in-plane oxygen coordinated to Mn) obey the I4/mmm symmetry.
The contribution of the oxygen to the total density is not sufficiently strong to indicate the
deviation from the higher symmetry.

If you look closely at the list of symmetry operators and their agreement factors, you will see
indeed that the symmetry operators of the space group I4/m are all at the beginning of the
list with the lowest agreement factors. But the difference in agreement factors between the
true opeators and the remaining operators of the I4/mmm space group is too small to be
detected by the program.

6. Refine the structure

Run Refine by double clicking the icon Refine

Refinement should immediately converge to an R-value around 5%
“Parameters → Atoms → Edit → Select all → Action → Edit/Define atoms → harmonic”
Change the type of displacement parameters of all atoms from isotropic to harmonic
Run Refine again
Refinement should converge to an R-value around 3.8%

This exercise demonstrates that the best approach to determining the space group is to
combine the information from diffraction pattern and the symmetry determination of
Superflip. In most cases the two approaches yield the same result. If the results differ, the
structure or data usually exhibit some non-standard features. It can be a strong
pseudosymmetry (as in this example), problem in the evaluation of systematic absences, or
twinning. The two approaches are complementary, and have different weaknesses. Their
combination gives the best result.

7. Additional exercises

a) Try to solve and refine the structure in I4/mmm. Apart from the high Rint, would you be
able to recognize that something is wrong with the structure model? What R-value do
you get?

b) For training purposes you can try to finish off the structure – check for the presence of
hydrogen atoms, define them and refine.

364 Jana2006 Cookbook, version October 2017

Example 9.5: 8rings
Very large structure – use of non-random starting point in Superflip

Revised: 12 February 2014

Input files: 8rings.hkl, 8rings.p4p. This is a huge organometallic structure with three eight-
membered and one three-membered ring containing chromium, nickel, iron and cobalt. Cell
volume 67312 Å3.

Refrence: G. F. S. Whitehead, F. Moro, G. A. Timco, W. Wernsdorfer, S. J. Teat and R. E. P.

Winpenny, Angewandte Chemie International Edition, 2013, 52, 9932-9935.

1. Load the data

Start Jana2006;
“File → Structure → New”;
Locate the input files, and double click on the green name 8rings in the right pane;
Select the radiobutton “known diffractometer formats”; NEXT;
Select the format Bruker-CCD; OK
OK to skip the warning about incomplete information;
Continue with the data import wizard, leaving all setting default, until the screen with the
information about the symmetry wizard
At this moment you loaded the basic crystallographic information and intensity data.

2. Determine the symmetry

When the reflection import finishes, you are automatically offered the symmetry
determination form. If you accidentally close this card, you can access the same functions
through the EditM50 icon, card Symmetry.
NEXT to start the wizard;
NEXT to confirm the default settings of the test;
OK to the information about no higher metric symmetry;
On the screen with Rint for different Laue groups make sure you understand Jana’s choice
for the 6/m Laue group.
NEXT;
On the screen with centering statistics make again sure that you understand the meaning of
the numbers and the choice Jana2006 made.
NEXT;
At this moment you are presented with the list of possible space groups. Based on the
analysis of systematic absences Jana suggests the space group P65. You may click the
button “Details” to get a detailed listing with the absent reflections. In this case there is little
doubt that the correct space group is indeed P61/P65.

3. Create refinement reflection file

Now you will filter and merge the raw reflection ata to produce the final reflection file ready
for solution and then refinement. In this case all settings can be left default:

365 Jana2006 Cookbook, version October 2017

 Answer YES to the question if you want to create refinement reflection file
Leave all settings default; NEXT; OK; NEXT; OK;
Accept the data block; FINISH

4. Structure Solution
As the next step the wizard suggests starting the structure solution. When you accept the
suggestion, a form with options for structure solution opens.

This is a large structure, and normalizing the intensities is most likely necessary.
Check the checkbox “Use local normalization”.

It is also necessary to give some composition. It is not used for solution but for interpretation. It
is sufficient to enter just a list of present atoms. Iron, cobalt and nickel will most probably be
indistinguishable in the solution, but we may enter them anyway, so that we do not have to add
them later during the refinement process. The same holds for hydrogen.
In the files Composition enter “Cr Fe Co Ni C H F N O”
Leave other settings default.
Click “Run solution”

Superflip starts and tries to solve the structure with default options for the charge flipping
algorithm. You may let the iteration run for a few hundred cycles and you will see that there are
no signs of convergence in the three indicators (R-value, total charge and peakiness).
There is still a chance that the structure is solved – sometimes the convergence is just not
visible. We will interrupt the iteration and check the result.
In the window of Jana click the button “Interrupt”

Superflip finishes the calculation and the log-file opens in the main window of Jana. Check the
symmetry analysis. The final part will look like this:
-------------------------------------------------
Space group derived from the symmetry operations:
-------------------------------------------------
HM symbol: P1
Hall symbol: P 1
Fingerprint: 33001041a0 (0,0,0)
Symmetry operations:
1: x1 x2 x3

It is most unlikely that the structure really has symmetry P1. Together with the lack of indicators
of convergence this means that the structure is not solved.

The structure contains moderately heavy atoms. It is thus sensible to start the iteration with
Patterson superposition map instead of the random map.
Click the button “Close” in the upper row to close the viewer of the log file;
In the selection list labeled “Starting model” select the radiobutton “Patterson superposition
map”:

Run the solution again

366 Jana2006 Cookbook, version October 2017

Typically the iteration converges within a few hundred cycles. The space group determination
part of the log file looks like this:
-------------------------------------------------
Space group derived from the symmetry operations:
-------------------------------------------------
HM symbol: P61
Hall symbol: P 61
Fingerprint: 3300163Xx01 (0,0,0)
Symmetry operations:
1: x1 x2 x3
6+_1(0,0,1): x1-x2 x1 1/6+x3
3+_1(0,0,1): -x2 x1-x2 1/3+x3
2_1(0,0,1): -x1 -x2 1/2+x3
3-_2(0,0,1): -x1+x2 -x1 2/3+x3
6-_5(0,0,1): x2 -x1+x2 5/6+x3

This time the solution has the correct symmetry and it is a strong indication that the structure is
solved.
Close the log-file viewer;
Wait until Jana performs the peak search (be patient, it is a large structure…)
Click the button “Draw 3D map” to view the resulting electron density
The density and the interpreted model are shown overlaid in VESTA. The structure is large and it
is not easy to understand the density. Nevertheless, after a while it becomes clear that the
structure is correctly solved and the main expected motifs are well visible. It is also visible that
some of the peaks present in the electron density were not interpreted as atoms by the peak-
seaching routine. These peaks are saved as “Fourier maxima” in Jana and you may change them
to regular atoms when you go to Parameters → Atoms → New → Peaks from the last Fourier
calculation. To see which maxima are interesting, we will plot the strongest peaks together with
the structure:

Close VESTA;
Click the button “Draw structure”;
In the next dialog click “Add Fourier peaks”;
A dialog with a long list of maxima opens. First about 100 maxima are deselected, because they
represent existing atoms. Then a large number of maxima are selected (typed in red).
OK to include all maxima;

The structure-plotting program (typically Diamond) opens and you may view the structure.The
added maxima are shown as the lightest element from the element list. In this case it is
hydrogen. You can see that a large number of these maxima are just artifacts close to the heavy
atoms, but among the added maxima you can also find most of the missing atoms. This
illustrates the fact that often the raw solution – the electron density – is more complete and
informative than the structure interpretation, and it is worth checking the electron density after
the solution, if the interpreted structure appears incomplete or otherwise wrong.

Here is for reference the image of the two chemically different entities found in the structure:

367 Jana2006 Cookbook, version October 2017

The three-ring with mixed iron-cobalt sites:

The eight-ring with chromium and nickel:

368 Jana2006 Cookbook, version October 2017

Example 10.1: AgSbS
Refinement of anharmonic ADPs in Jana2006.
Chemical formula: Ag3SbS3
Reference: F. Laufek, J. Sejkora & M. Dušek (2010). Journal of Geosciences, 55, 161–167
Single crystal data measured with Oxford Diffraction four-circle diffractometer
Input files: agsbs.hkl, agsbs.cif_od
Frame scaling, absorption correction: done with software of diffractometer

In this example we shall need installation of VESTA for visualization of electron density
associated with anharmonic ADP. VESTA is supplied as a zip archive which can be unpacked to
arbitrary location (with simple path – not to Desktop). Then the path to VESTA.exe must be
defined in “Parameters → Programs”.

1. Creating new jobname

Start Jana2006
“File → Structure → New” opens a file manager
Left pane: locate the directory with input files
Right pane: double-click agsbs
Right pane detects possible Jana files and shows one jobname for each group of files

2. Import Wizard
Select “Known diffractometer formats”; NEXT
Select “Oxford Diffraction – CCD”; NEXT
Select “Input from “cif_od” file”; OK
The used diffractometer software splits reflection file and another information (cell
parameters, wavelength …). There are several output formats for both cases.
Leave all settings unchanged; NEXT
Leave all settings unchanged; NEXT
The program reads 11948 reflections from hkl file
For absorption correction select “None or done before importing”; NEXT
FINISH

3. Data RepositoryOK;
Each line in data repository window corresponds to one data set converted to Jana format.
In our case we use only one data set.
OK; YES to accept the data set

4. Symmetry Wizard
[On the screen: “Tolerances for crystal system recognition”.]
Symmetry wizard can be started separately by “File → Reflection file → Make space group
test.
Leave all settings default; NEXT
[On the screen: Select Laue symmetry]

369 Jana2006 Cookbook, version October 2017

 Select Trigonal -3m1 Laue symmetry; NEXT
Select R-obverse cell; NEXT
[On the screen: Select space group]
Select R3c; NEXT
Selecting the non-centrosymmetric space group is a shortcut. It will be explained at
“Verification of structure solution”.
Accept the space group transformed into the original cell; FINISH
Symmetry is saved in file agsbs.M50.

5. Creating refinement reflection file

In this step program creates file agsbs.M90 containing the data set merged by symmetry
and with discarded forbidden reflections. M90 will be used for refinement.
NEXT to confirm threshold 3sigma
[On the screen: settings for merging of reflections]
Leave all settings default; NEXT
The program reports Rint 1.88% for 430 observed reflections merged from 3577 observed
reflections.
Accept the new data block and calculate coverage; FINISH

6. Structure Solution Wizard

[On the screen: window of Structure solution].
In “Formula” textbox type list of chemical elements: Ag Sb S
Select “Use Superflip”, “Peaks from Jana2006”;
Select “Use a specific random seed” and use “222”
Fixing the random seed to 222 guarantees for most processors that the Superflip returns
exactly the same results like described in this cookbook. Otherwise each run of Superflip may
return different equivalent origin. For normal work random seed should be automatic
because the convergence of charge flipping is not guaranteed for every value of random
seed.
Leave other settings default; OK
[On the screen: window of Superflip replaced after reaching the convergence with listing of
Superflip]
Superflip converges (after noise suppression) with R value (after noise suppression) 16%.
Press CLOSE to leave the listing
Press “Accept the result”

7. Verification of the structure solution

Start EditM50, go to page Composition, enter 6 for “Formula units”, press “Formula from
M40”
It displays formula Ag3 Sb S3 (for Z=6). This is the same as expected composition.
[On the screen: Information about formula]
OK; YES for updating the formula.
In next steps we shall verify with Superflip that centre of symmetry is not present.
Start “File → Structure → Save As”
Save the structure with jobname “testspg”
YES to continue with the new structure
Start “Run → Solution”
Select “Use Superflip” and press button “Change the space group”

370 Jana2006 Cookbook, version October 2017

 For the “Space group” type R-3c; OK; YES to rewrite M50
Follow steps for creating the refinement reflection file
Merging the data yields Rint 2.11% and the program returns to the structure solution wizard
[On the screen: Structure solution wizard]
OK to start structure solution
[On the screen: listing of Superflip]
At the end of the listing we can see unsatisfactory agreement factor for the symmetry
operator “2”. This indicates the space group is probably wrong.
Press “Accept the result”
Start EditM50, go to page Composition, enter 6 for “Formula units”, press “Formula from
M40”
It displays formula Ag0 Sb S6 (for Z=6) which is also wrong
Return to the basic window of Jana2006
Start “File → Structure → History” and open the previous structure (with space group R3c)

8. Refinement
[On the screen: basic window of Jana]
Right-click the icon of Refine.
[On the screen: refinement options]
Define 100 of cycles, damping factor 0.5; OK
Choose “YES+START”
Refinement converges with R value about 12%
Start “Edit atoms”
[On the screen: list of atoms]
Press “Select all” and “Action → Edit/define”
Choose “harmonic” for ADP parameters. This defines anisotropic temperature parameters
for all atoms. OK, OK, Yes to save changes
Double-click the icon Refine.
Refinement converges with R value about 4%.

9. Extinction correction
The results seem to be OK, but:
[On the screen: basic window of Jana2006]
Open listing of Refine (by “Edit/View → View of Refine”)
Press “Go to” and select “Statistics Fo, sin(th)/lambda after refinement”
R value should decrease with increasing intensity or with lowering the diffraction angle.
Here the R value for the strongest reflections as well as for the low angle reflections is higher
than in the previous shell. This suggests that extinction correction is required.
Close the refinement listing
“Parameters → Extinction”
Select “Isotropic”; OK
Refinement of extinction coefficient is enabled by default
Double-click Refine and refine the final structure
Refinement converges with R factor 3.67 %.
Check the extinction coefficient through “Parameters → Extinction”
View listing of Refine; Go to “Statistics”; check improved R value for strongest reflections.

371 Jana2006 Cookbook, version October 2017

10. Anharmonic refinement
Run Contour, start “New plot” – “Calculate new ones” +” Draw maps as calculated”.
Here no previous map exists and therefore “Calculate new ones” will be disabled
[On the screen: Fourier commands]
In Basic thumbnail select “F(obs)-F(calc) – difference Fourier”; leave other settings default
In Scope thumbnail define the vicinity of Ag1 according to the following screenshot; OK

[On the screen: contour plot of the first section]

In “Contours” define both positive and negative contours to 0.1 e-/Å3;
In “Atoms edit” add Ag1;
Examine the sections using "M+" and "M-"

372 Jana2006 Cookbook, version October 2017

We can see residua of positive density in the vicinity of the Ag atom.

Quit Contour

Run “Edit atoms” and select Ag1.

“Action → Edit/define atoms”: check “Anharmonic ADP”, third order;
Run Refine
Refinement converges with R factor 1.14 %.
In Contour draw the same difference Fourier section as previous (calculate new maps). Use
the same limits for contours.

373 Jana2006 Cookbook, version October 2017

 The residua have significantly decreased.

11. Density modeling using the joint probability density function

Run Contour, start “New plot” – select j.p.d.f.
[On the screen: Select atoms and basic parameters for j.p.d.f]
Select Ag1 and S1; Next
[On the screen: Plane/Volume]
For 1st atom use Ag1; for the 2nd and 3rd atom use difference to 1st 1 0 0 and 0 1 0,
respectively;
Define scope of the section: 6 6 6 and “1st point to” 3 4 2.5;
Use cutoff distance for atoms 4 Å; NEXT

374 Jana2006 Cookbook, version October 2017

[On the screen: Select individual atoms for j.p.d.f]
Select Ag1, and all symmetry related sulphur atoms close to Ag1 and note their symmetry
codes (i.g. S1#s5c3t0,-1,-1); NEXT

In “Atoms Edit” of Contour define the Ag1 atom and all four S1 atoms using their symmetry
codes.
Press “Run 3d maps”
Jana2006 will call an external program for drawing isosurfaces, previously defined in
“Tools → Programs”, in our case VESTA
Setting drawing options:
In “Edit → Bonds” define bond between Ag and S; maximal distance 3.15 Å,
In “Objects → Structural model” select “Stick”,
In “Properties” check “Do not show unit cell”.

375 Jana2006 Cookbook, version October 2017

 The application of non-harmonic parameters led to the triangular shape of the electron
density centered in Ag1 position. The electron density is extended in directions between of
Ag-S bonds.

12. Drawing one particle potential curve

Sulphur and Silver form spirals defined by short Ag-S bonds. The possible exchange paths for
silver atoms within and between the spirals can be tested with help of one particle potential
along a proposed path. One of proposed paths is shown in the following figure where it goes
along sides of the indicated triangle:

376 Jana2006 Cookbook, version October 2017

Run Contour, start “New plot” – select j.p.d.f.
[On the screen: Select atoms and basic parameters for j.p.d.f]
Select Ag1; Next
[On the screen: Plane/Volume]
Define the displayed volume according to the following screenshot; OK

Select Ag1#s3c2, Ag1#s5c3t0,-1,0 and Ag1; NEXT

In “Curves” “Use mouse to define polyline” and draw the line Ag1(a)—P—Ag1(b)—P—Ag1(c)—
P—Ag1(a) by mouse (2nd mouse button quits the selection). You can also fill in these points
directly in the textboxes of the “Curves” dialogue. Fractional coordinates of the point P are
approximately 0.267 -0.63 0.235. Ag atoms should be written with corresponding symmetry
codes, i.e. Ag1#s3c2.

377 Jana2006 Cookbook, version October 2017

In the first plot we can see an electron density curve. The maximal electron density is located in
Ag1 atoms.

Press “Continue”
Adjust “y-Scale”: Min and Max in meV: 0 600;

The resulting curve is one particle potential path. The peaks represent the potential barrier
between Ag1 atoms when the exchange path runs through the point P. These values suggest
that the silver transfer within and between neighboring Ag–S–Ag spirals is possible.

378 Jana2006 Cookbook, version October 2017

379 Jana2006 Cookbook, version October 2017
380 Jana2006 Cookbook, version October 2017
Example 10.2: Oxa
multipole refinement

X-ray data from an oxalic acid sample measured at 110K to a high resolution limit. The crystal
contains also crystal water: C2O4H2.2H2O

Input files: oxa.m40, oxa.m50, oxa.m90 and oxa.m95 – Jana files from a conventional
refinement

25. Drawing density map based on the regular refinement

[On the screen: Basic Jana window]
Start Jana2006
“File → Structure → Open”: open structure “oxa”
This is result of conventional structure refinement
Run refinement
Refinement converges in 1 cycle to R(obs)=2.74%
[On the screen: Basic Jana window]
Double-click the icon Contour; Press the button “New plot”
Select “Draw a general section”; OK;
[On the screen: Fourier commands]
In “Basic” page for “Map type” select “F(obs) – Fourier”
clear “Omit non-matching reflections”
select “Use weighting of reflections”
clear “Apply sin(th)/lambda limit”
In “Scope” page select “automatically”, “Independent parallelepiped” and “Use default map
orientation”
Change the step to 0.05 Å
OK
[On the screen: Plane/Volume]
This form defines the general section to be drawn from the calculated map.
Define: 1st atom C1; 2nd atom C1#s3; 3rd atom O1
#s3 encodes symmetry operation (the third one in the list in EditM50)
Set the “Interpolation step” to 0.05 Å, “Scope of section” 6 6 0; 1st point to 2 3 0
OK;
After calculation the density section through oxalic acid molecule will appear:

381 Jana2006 Cookbook, version October 2017

Atomic positions can be included into the map by clicking the pointer at different locations
and selecting atoms close to the given location. The following atomic position should be
selected: C1, O1, H1, O2, C1#s3, O1#s3, H1#s3; O2#s3.

Quit; Save it as Plane#1; OK

[On the screen: Basic Jana window]
Now we shall visualize the difference map.
Double-click the icon Contour; Press the button: ”New plot”
Select “Calculate new ones” and “Draw a general section”; OK;
[On the screen: Fourier commands]
In “Basic” page for “Map type” select “F(obs)-F(calc) – difference Fourier”
Keep all remaining settings unchanged; OK;
[On the screen: Plane/Volume]
Select “Use old”; OK
[On the screen: Difference Fourier map through the oxalic acid]
Press the button “Atoms ON”

382 Jana2006 Cookbook, version October 2017

 Press the button “Contours” to see contour interval. The value should be about 0.05 e-/Å3.
The positive density clouds are in vicinities of possible bonds. This indicates that the
multipole model can reflect real charge densities in this crystal.
Quit Contour

26. Kappa refinement

[On the screen: Basic Jana window]
We shall keep backup copy of the conventional refinement. Refinement will now continue
with new job name.
Start “File → Structure → Save As”
Save structure as “Oxa-Kappa”
Continue with the new structure
[On the screen: Basic Jana window]
Double-click the icon EditM50
In page “Composition” define multipoles “MOLLY”;
In page “Multipole parameters” define orbital populations; for O and C atoms keep the
default values; for H atom select “Slater form factor” with the default values for H-N and
H-Dzeta;
OK; YES to rewrite M50
[On the screen: Basic Jana window]
Double click the icon “Edit atoms”;
Select all atoms; Choose “Action → Edit/define” or double click on the selected atoms;
[On the screen: Atom edit]
In the column “Multipole parameters” select “kappa”; OK; OK;
Yes to rewrite the changed files
[On the screen: Basic Jana window]
Right click on the icon “Refine”;
In the “Various” page press “Equations” and define the following equations:
pv[H3]=pv[H2]
kappa[H2]=kappa[H1]
kappa[H3]=kappa[H1]
OK; OK; Yes+start

383 Jana2006 Cookbook, version October 2017

 Refinement converges to R(obs)=2.69%
The kappa refinement made only small improvement. This is in accordance with the
difference Fourier map in which strong asphericity has been visible.

27. Distance and angular restrains for H atom

Way how hydrogen atoms are described is a delicate point for charge density refinement.
Generally hydrogen electron can be related to any reference point. In order to have
reasonable deformation maps it is worthwhile to fix the reference positions of hydrogen
atoms close to its ideal values. For the oxalic acid we can use the distance and angular
restrains to values refined against neutron diffraction data. For distance H1-O1 this value is
1.07 Å with angle H1-O1-C1 113o and for O3-H2 and O3-H3 0.97 Å with angle H2-O3-H3
about 106o.
[On the screen: Basic Jana window]
Right click on the icon “Refine”;
In the page “Various” define the following “Distance restrains”:
O1 H1; Value 1.07; s.u. 0.001;
O3 H2; Value 0.97; s.u. 0.001;
O3 H3; Value 0.97; s.u. 0.001;
OK;
In the page “Various” define the following “Angle restrains”:
H1 O1 C1; Value 113; s.u. 0.01;
H2 O3 H3; Value 106; s.u. 0.01;
OK
In the page ”Various” define the following “Keep commands”:
ADP; Riding; Central O1; Extension 1.2; One hydrogen: H1;
ADP; Riding; Central O3; Extension 1.2; Two hydrogens: H2, H3;
OK; Yes+start;
Use all to “significant discrepancies”;
Refinement converges to R(obs)=2.74%

28. Multipole refinement

For defining of multipole parameters it is convenient to select a local coordinate systems for
each atom. The following figure shows a recommended choice for atoms C1, O1, O2 and H1:

384 Jana2006 Cookbook, version October 2017

The axis x is oriented along the bond from the central atom to one of non-hydrogen atoms
(here called "first atom"). Our choice of “first atom” for central atoms C1, O1, O2 and H1 is
O2, C1, C1 and O1, respectively. According to the definition in Jana2006, the second axis
(here y) is perpendicular to the axis x and it lies in the plane defined by the central, first and
second atom in such a way that the angle between the axis y and vector from the central
atom to the second one is smaller than 90o. Our choice of “second atom” for central atoms
C1, O1, O2 and H1 is O1, O2, C1 and C1, respectively. Note that second atoms need not
make bonds with the central atom.
The similar choice can also be used for the molecule of crystal water.
[On the screen: Basic Jana window]
In order to save the result of the kappa refinement a new structure will be opened:
Start “File → Structure → Save As”
[On the screen: File manager for saving structure]
In File manager focus the right-bottom textbox and define a new name for testing of the
kappa refinement i.e. “Oxa-Multipole”; OK
Yes to continue with the new structure
[On the screen: Basic Jana window]
Run “Edit atoms”;
Double click the first atom – C1; switch from “kappa” to “order”
Change multipole parameters to order 3 and define the coordinate system x → O2 and
y → O1.

Using the list box on the top left corner go to the next atom – O1.
Change multipole parameters to order 3 and define the coordinate system x → C1 and
y → O2. Go to the next atom – O2.
Change multipole parameters to order 3 and define the coordinate system x → C1 and
y → O1. Go to the next atom – O3.
Change multipole parameters to order 3 and define the coordinate system x → H2 and
y → H3. Go to the next atom – H1.
Change multipole parameters to order 1 and define the coordinate system x → O1 and
y → C1. Go to the next atom – H2.
Change multipole parameters to order 1 and define the coordinate system x → O3 and
y → H3. Go to the next atom – H3.
Change multipole parameters to order 1 and define the coordinate system x → O3 and
y → H2.
OK; OK; Yes to the question “Do you want rewrite changed files?”
[On the screen: Basic Jana window]
Right click on the icon “Refine”;
In the page “Basic” select the number of cycles 100 and damping factor 0.1;
In the page “Various” press “Fixed commands” and fix the following parameters:
385 Jana2006 Cookbook, version October 2017
 “Set individual”, P00[*], “Set to” 0;
“Individual”, kappa'[*];
[*] stands for all atoms. P000 wil be zero for all atoms while kappa’ will be fixed to the
current value.
OK; OK; YES+Start
Refinement converges to R(obs)=1.75%.

29. Multipole refinement including kappa' parameters

[On the screen: Basic Jana window]
Right click on the icon “Refine”;
In the “Various” fix the following parameters:
Delete or disable “Individual”, kappa'[*];
In the “Equation” define two equations for H atoms;
kappa'[H2]=kappa'[H1]
kappa'[H3]=kappa'[H1]
OK; OK; Yes+start;
Refinement converges to R(obs)=1.69%.
It may happen that at this point R value is higher, around 1.75%. That means water
molecule is in a false minimum. In such case: change number of multipoles for O3, H2 and
H3 to zero; fix kappa’; repeat refinement; change number of multipoles back to 3,1,1 (for
O3, H2, H3); repeat refinement; release kappa’, repeat refinement

30. Fourier maps

[On the screen: Basic Jana window]
Double-click the icon Contour; Press the button “New plot”
Select “Calculate new map” and “Draw a general section”;OK;
[On the screen: Fourier commands]
In “Basic” use default setting i.e. Map type F(obs)-F(calc) – difference Fourier;
Keep all remaining parameters unchanged; OK;
[On the screen: Plane/Volume]
Select “Use old”;OK
[On the screen: Difference Fourier map through the oxalic acid]
Press the button “Atoms ON”.
Press the button “Contours” and modify the positive and negative contours to 0.05 e-/Å3

The map does not show any clear positive or negative maxima which means that the model
explain the experimental density adequately.

386 Jana2006 Cookbook, version October 2017

31. Interpretation of multipole parameters – map visualization
[On the screen: Basic Jana window]
Double-click the icon Contour; Instead of “Fourier maps” select ”Charge density maps”;OK
[On the screen: Define atoms and basic parameters for charge density maps]
Select all atoms; Calculate “deformation density”;NEXT;
[On the screen: Plane/Volume]
Select “Use old”; NEXT
[On the screen: Select individual atoms for charge density]
Select all atoms; NEXT
Press the button “Atoms ON”;
After a short calculation the following map should appear.

32. Topological analysis – critical points

[On the screen: Basic Jana window]
Run “Tools → Topological analysis → Find critical points”
[On the screen: Define parametrs for CP search]
Select “Type of search” “between all bonds within interval”;
For all other parameters keep in default values;
OK;
[On the screen: Select atoms for distance check]
Select all; OK
[On the screen: Basic Jana window]
Run “Edit/View → View of CP report”;
[On the screen: Viewer for critical points]
The listing contain information about critical point along all important bonds in the
molecule.

Fractional coordinates
Bond ρ Laplacian Hess eigenvalues Ellipsity
x y z

387 Jana2006 Cookbook, version October 2017

 C1-C1[-x,-y,- - -
0.0000 0.0000 0.0000 1.77 -13.55 11.08 0.23
z] 13.59 11.04
C1-O1 - -
0.0011 0.0187 0.0875 2.50 -31.96 11.41 0.13
22.99 20.38
C1-O2 - - -
0.1228 0.0465 2.97 -36.60 19.71 0.17
0.1076 30.34 25.97
O1-H1 - - -
0.0331 0.2065 1.62 -22.14 22.66 0.01
0.0015 22.47 22.34
O3…H1 0.0041 0.0592 0.2493 0.54 5.16 -3.86 -3.36 12.38 0.15
O3-H2 - -
0.0499 0.1790 0.3755 2.21 -49.22 23.40 0.02
36.74 35.88
O3-H3 - - - -
0.3421 2.30 -61.79 24.39 0.03
0.1228 0.0092 43.69 42.49

33. Topological analysis – Basin integration

[On the screen: Basic Jana window]
Run “Tools → Topological analysis → Basin integration”
[On the screen: Select atoms for basic integration]
Select all; OK
[On the screen: Define parameters for integration]
Use default values; OK
Calculation can take some time …
Run “Edit/View → View of Inb report”;
The integrated values for individual atoms should be:

Number of
Atom Charge
electrons
C1 5.147 0.853
O1 9.124 -1.124
O2 8.773 -0.773
O3 9.378 -1.378
H1 0.396 0.604
H2 0.407 0.593
H3 0.419 0.581

388 Jana2006 Cookbook, version October 2017

Example 11.1: bezcox
Composite structure

Sr1.272CoO3
Revised: 28 April 2014
Single crystal data measured with Oxford Diffraction four-circle diffractometer
Input files: bezcox.hkl, bezcox.cif_od
Frame scaling, absorption correction: done by diffractometer software

1. Creating new jobname

Start Jana2006
“File → Structure → New” opens a file manager
Left pane: locate directory with input files
Right pane: double-click bezcox

2. Import Wizard
Select “Known diffractometer formats”; NEXT
Select “Oxford Diffraction – CCD”; NEXT
[On the screen: “Define basic input file:"]
Select “Input from cif_od file”; file name “bezcox.cif_od”; OK
[On the screen: “Complete/Correct experimental parameters”]
Leave all settings unchanged; NEXT
[On the screen: “Define the reference cell/split by twinning”]
Leave all settings unchanged; NEXT
For absorption correction select “None or done before importing”; NEXT
FINISH

3. Data Repository
OK; YES to accept the data set
[On the screen: “Information about starting symmetry wizard”]
NEXT

4. Symmetry Wizard
[On the screen: “Tolerances for crystal system recognition”.]
Symmetry wizard can be started separately by “File → Reflection file → Make space group
test”.
Leave all settings default; NEXT
[On the screen: Select Laue symmetry]
Select Trigonal -3m1 Laue symmetry; NEXT
[On the screen: Select cell centering]
In the list we can see that R-reverse centering is almost fulfilled.
Click on “Details” and examine strongest reflections contradicting R-reverse centering
The most of printed reflections follow the rule -h+k+l=3n. Similarly, reflections contradicting
R-obverse centering follow the rule h-k+l=3n. This can be a sign of obverse/reverse twinning.
Return back to the page with tolerances and check there “Introduce twin laws in case of
subgroup”

389 Jana2006 Cookbook, version October 2017

 Follow steps of the symmetry wizard, select -3m1 Laue symmetry, NEXT
Now, only 128 reflections contradicts the R (obverse or reverse) centering
Select R-reverse; NEXT
Before introducing the twinning, R-reverse centering was better fulfilled. This indicates that
the major twin domain has R-reverse centering.
Select R-3m; NEXT
[On the screen: Select superspace group]
R-3m(00g)0s
[On the screen: Final step of the superspace test]
Accept the superspace group in the standard setting: R-3m(00g)0s;
Accept twinning matrices induced by the space group test; FINISH; OK
Continue with creation of reflection file with default options;
[On the screen: Structure solution]
Press “Quit”;

5. Define W-matrix
Using Tools → Reciprocal space viewer we can clearly distinguish the composite subsystems.
This compound has the simplest W matrix, which only exchanges the “l” and “m” index.
Than means that “h k l m” are reflections of the first subsystem while “h k m l” are
reflections of the second subsystem.
Open “Edit M50”, change number of composite parts to 2 and define W-matrix for the
second subsystem: (1 0 0 0 | 0 1 0 0 | 0 0 0 1 | 0 0 1 0 )

“Induced cell paramemeters” are parameters of the second subsystem, as follow from the W
matrix

6. Structure Solution
Double-click the icon “Structure solution”.
[On the screen: window of Structure solution]
In “Formula” textbox type list of chemical elements for Sr1.272CoO3 : Sr Co O
Select “use Superflip”; “Peaks from Jana2006 but first run Fourier”
Select “Use a specific random seed” and type value “111”
Press “Run solution”;
The option “Peak from Jana2006 but first run Fourier” in necessary for proper interpretation
of Fourier map for composite structure because we need to look for bot subsystems
Superflip converges (after noise suppression) with R value 20%. It confirms the symmetry
R-3m(00g)0s.
Fixing the random seed to 111 guarantees for most processors that the Superflip returns
exactly the same results like described in this cookbook. For this example finding proper
solution would require many attempts with Superflip.
Press “Quit” to leave the structure solution wizard

7. Verification of the structure solution

Double click on the icon “Edit atoms”;
390 Jana2006 Cookbook, version October 2017
Select oxygen atoms if present and delete them;
At this stage found oxygen positions are not reliable;
Double click on the Co1 atom and check its coordinates;
Co1 0.333333 0.666667 0.166667 - in the first subsystem
Move to the Co2 atom;
Co2 0.000000 0.333333 0.250000 – in the second subsystem
The solution program cannot distinguish at this stage between chemical types of heavy
atoms because both atoms are placed in different subsystem and we cannot compare their
electron densities directly.
OK;OK; Yes to save the changes;

Double click on the icon “Contour”;

Press “New plot”; Select “Calculate new ones” and “Draw maps as calculated”; OK;
[On the screen: window of Fourier commands]
On the page “Basic” define Composite part no. 1;
Map type “F(obs) – Fourier”;
On the page “Scope” select the central point Co1; select the x3-x4 section; scope 1 1 2; OK

391 Jana2006 Cookbook, version October 2017

Repeat the procedure for Co2;

x3-x4 section for Co1 x3-x4 section for Co2

Note: Both maps are summed over x1 and x2.

Press “New plot”; Select “Calculate new ones” and “Draw maps as calculated”; OK;
[On the screen: window of Fourier commands]
On the page “Basic” define Composite part no. 2;
Repeat map calculations for both atoms.

392 Jana2006 Cookbook, version October 2017

 x3-x4 section for Co1 x3-x4 section for Co2

Start “Parameters → Twin fractions” and change the twin fraction to 0.5
Activate refinement of twin fraction
In Refinement options change number of cycles to 100
Run refinement
Refinement converges with R values 34%, 24%, 41%, 58%, 67% and 97%
Start “Edit/View → View of Refine” and read information following the header “Composite
part #2”.
Here we can see unit cell and symmetry of the second composite subsystem. Note that the
second subsystem has also different centering because centering vectors are also
transformed by W matrix.

393 Jana2006 Cookbook, version October 2017

 Double click on the icon “Edit atoms”; Double click on the atom Co2;
The “Edit” page shows that the parameter Uiso is negative. It indicates that atom type is Sr.
Go the “Define” page and change the atom type to Sr and rename it to Sr1;
Run refinement
Refinement converges with R values 32%,21%, 40%, 58%, 70% and 97% and the isotropic
ADP parameters become positive.

8. Making results compatible with Acta Cryst. (1999), B55, 841-848

Start “Tools → Transformation → Origin shift” and move origin to the position of Co1.
Position of Sr1 changes to -1/3, -1/3, 1/4.
Start Edit atoms, select Sr1, Action: Transform selected atoms
Select symmetry operator –x2 x1-x2 x3 x4
Run Refinement
R factors of the previous refinement are exactly repeated
Now we are compatible with published positions of Co1 and Sr1.

9. Oxygen
Run refinement to obtain input for Fourier
Run difference Fourier for composite part No 1, with Scope=”automatically”, default map
orientation, independent parallelepiped
We use composite part 1, because we expect oxygen belonging to Co coordination
YES to start procedure for including atoms
Locate the strongest maximum with reasonable distance to Co (probably Max2)
Include this maximum as O1 in position about 0.17, 0.04, 0.5
Start Contour and plot section x3-x4 through O1, scope 1 1 2, difference Fourier, using
summation of x1, x2 over 1 Å;
394 Jana2006 Cookbook, version October 2017
 section x3-x4 of O1

The section reveals O1 should be described with crenel function, x40=0.75, delta=0.5

Plot section x1-x2, difference Fourier, x4 from 0 to 1, scope 5 5 0, step 0.5A, center 0 0 0.5
The section x1-x2 with center at (0,0,1/2) with scope about 5 Å as a function x4 shows the
character of the modulation

395 Jana2006 Cookbook, version October 2017

 Run “Movie” to see the discontinuous character of the modulation.
The three oxygen positions are generated by the three fold axis. From the above figure on
epossibility for oxygen position is about (0.157,0.157,0.5) (as indicated by the arrow) with
crenel of length 0.5 with x4 center 0.25.
Start Edit atoms and double-click O1
Define the crenel function for O1 and delete its position modulation wave
In page Edit press Occupancy and set x40=0.25 and delta=0.5 (page “Edit”).
In page Edit modify the atomic position of O1 to the one found in the difference Fourier and
reset size occupancy (it should be 0.5).
Finally in the page “Define” define type of modulation functions – Legendre polynomials.
Run refinement
Refinement converges with R values 26%,13%, 36%, 58%, 66% and 88%
Set harmonic ADP for all atoms
Repeat refinement
R values are almost unchanged

10. More modulation waves

Add one by one more position and ADP modulation waves.
Co: 4 for position, 4 for ADP
O1: 4 for position, 2 for ADP
Sr1: 4 for position, 4 for ADP
The last refinement will converge with R values 4.3%,2.5%, 4.2%, 9.2%, 21.5% and 59%
With fourth order satellites it is doubtful if we could achieve better results because they are
almost completely unobserved. However, the fit for 2nd and 3rd order satellites is not yet
satisfactory

11. Sawtooth for Sr1

Start Contour and plot x3-x4 section (observed and difference Fourier) for Co1, O1 and x2-
x4 section for Sr1. Use 1 Å summation. Do not forget to switch composite subsystem in the
Basic page of Fourier commands.

396 Jana2006 Cookbook, version October 2017

 x3-x4 section for Co1 difference section, contour step 0.2

x3-x4 section for O1 difference section, contour step 0.2

397 Jana2006 Cookbook, version October 2017

 x2-x4 section for Sr1 difference section, contour step 0.2

The largest different peaks are around Sr1 position. Now we shall attempt to describe
modulation function of Sr1 with Lagrange polynomials, with crenel defined at x40=0 with
delta=1

Make backup copy (File → Structure → Save As)

Start Edit atoms
Remove all ADP and position modulation waves of Sr1 except the first position modulation
wave;
Activate in the Crenel for Sr1 and the the edit page set x40=0 and delta=1. Keep x40 and
delta fixed
In the page “Define” Parameters activate Legendre polynomials
In refinement commands set Damping factor 0.1
Run refinement
During the refinement change Damping factor to 0.5
Refinement converges with R values 10.3%, 4.0%, 9.3%, 37.7%, 46% and 53%
Add one by one position and ADP modulation waves for Sr1 up to three position waves and
three ADP waves.
The last refinement will converge with R values 3.7%, 1.9%, 3.6%, 8.2%, 18.7% and 55%. This
is significant improvement comparing with the structure model without sawtooth function.

Anharmonic ADP for Co1

“Start Contour → New plot → Calculate new ones”

In Basic page select Difference Fourier and Composite part no. 1
In Scope select central point Co1, section x1-x2, scope 6 6 1
OK

398 Jana2006 Cookbook, version October 2017

In Contour make summation only along x3 (not x4)

x1-x2 section through Co1, contour step 0.05, x4=0

399 Jana2006 Cookbook, version October 2017

x1-x2 section through Co1, contour step 0.05, x4=0.26

The neighborhood of Co1 is changing from octahedron to trigonal prisma (check it with
drawing program). The trigonal coordination induces a slightly triangular shape of ADP.

However, the region in x4 in which the effect is detectable is very narrow and the number of
modulation waves for ADP’s of the order 3 would be very large to describe such abrupt
change.

The atom Co1 should therefore be split into two positions by crenel – the first one Co1
(x40=0, delta=0.45) and the second one Co1’ (x40=0.25, delta 0.05). After adjusting of
occupancies to ai=1/6 for both atoms the two Legendre waves for position and harmonic
ADP Co1 should be used.

For the Co1’ only average parameters for position and ADP up to the third order.

Final R values : 3.2%, 1.8%, 3.0%, 7.3%, 13.9% and 55%

400 Jana2006 Cookbook, version October 2017

Example 12.1: Ba5Co5
Simple antiferromagnetic magnetic structure (k = 0)
Revised: 23 October 2017

Ba5Co5ClO13
Powder data measured with the ILL neutron source at 1.4 K
For more details: Mentré, O., Kauffmann, M., Ehora, G., Daviero-Minaud, S., Abraham F. &
Roussel, P. (2008). Solid State Sciences 10, 471-475.

Input files: Ba5Co5.dat (powder profile data)

Ba5Co5_nuclear.cif (nuclear structure)

In order to facilitate the work with magnetic structures, Jana2006 offers special wizards and
protects the basic file - the parent structure. However, for experienced users of Jana, all
operations offered by wizards can be also done manually by starting the tools directly from the
basic user interface. The standard work with magnetic structures based on powder data follows
this scheme:

Reading of powder profile and parent structure model

Wizard for le Bail refinement of powder profile parameters

Leaving this wizard automatically fixes all profile parameters except
the background

Wizard for Rietveld refinement of the parent structure

Leaving this wizard automatically locks the parent structure in order
to prevent its accidental refinement

Wizard for selection of the magnetic symmetry

Magnetic structure refinement using the selected magnetic symmetry

This runs under different jobname in order to keep the parent structure
unchanged. Each magnetic symmetry should be tested under a unique
jobname.

Loading the parent structure for testing another magnetic

symmetry

401 Jana2006 Cookbook, version October 2017

1. Creating new jobname
Start Jana2006
“File → Structure → New” opens a file manager
Left pane: locate directory with input files
Right pane: double-click Ba5Co5
[On the screen: Specify type of the file to be imported]
Select “Magnetic parent structure: nuclear model from CIF”; NEXT
[On the screen: Select input CIF file]
Right pane: select (by the button with arrow) the input file Ba5Co5_nuclear.cif; OK

2. Define magnetic propagation vector and form factors

[On the screen: Define magnetic propagation vector and form factors]
For Number of non-zero magnetic propagation vectors select 0
For “Atom type” select “Co”; check “Use as a magnetic atom”
For “Magnetic formfactor <j0>” select Co4+; NEXT; Finish
The parent structure has been created

3. Import the reflection file

[On the screen: Specify type of the file to be imported]
Select “Powder data - Various CW formats”; NEXT
Check that the name of the input file is “Ba5Co5.dat” and the format is “Free format of 2th,
I, [sig(I)]”; NEXT
[On the screen: Complete/correct experimental parameters];
For “Wave length” type 2.52; NEXT; Finish
[On the screen: Data repository]
OK; YES to “Do you want to accept made changes?”

Refining of profile parameters by le Bail technique.

[On the screen: Information]
NEXT;
[On the screen: Refinement of the powder profile by the le Bail algorithm]
“Show powder profile”
As common for powder neutron data the background is strongly varying with the diffraction
angle. It is recommended to estimate it manually and use just a few additional parameters.
Manual background: “Create new”; Use default parameters; OK
The automatic procedure has created a first estimation of the manual background, which
can be modified manually: the points can be dragged by the left mouse button; dragging a
point to the position of neighboring point deletes the point; clicking between two points
creates a new point.
Using "X+", "Y+", enlarge the profile plot
Modify background point at high and low angles as indicated in the next figures:

402 Jana2006 Cookbook, version October 2017

“QUIT”; Select “Accept the new manual background” and “Reset polynomial coefficients to
zeros”; OK

403 Jana2006 Cookbook, version October 2017

 This means that the polynomial background will be refined and combined with the manual
background
“Edit profile parameters”;
[On the screen: Powder options]
In the page “Cell” activate refinement of a and c parameters;
In the page “Profile” check if starting value of GW is 50 and activate its refinement;
In the page “Corrections” activate refinement of “shift” parameter; OK; YES to save changes
“Run Refine”;
The refinement should converge to Rp~2.7%.
“Edit profile parameters”;
[On the screen: Powder options]
In the page “Profile” change activate refinement of GU and GV; OK; YES to save changes
“Run Refine”;
The refinement should converge to Rp~2.1%.
"Show powder profile" to see the fit of the calculated profile; QUIT
NEXT;
YES to “Do you really want to leave the form for le Bail refinement?”
Now we will refine the scale parameter for the nuclear structure
[On the screen: Refinement of the nuclear structure by the Rietveld method]
“Edit refinement commands”; On the page “Various” select “Fixed commands;
Check that the fixed commands “fixed all *” is present if not select “All parameters” and fill
into the editing window “Atoms/parameters” by the wild character “*” which means all
atoms
Press "Add";
The command will appear in the panel as “fixed all *”
OK; OK;
“Run Refine”;
The refinement should converge to Rp~3.3%.
NEXT;
YES to “Do you really want to leave the form for Rietveld refinement?”
NEXT to accept the Information

4. Testing different irreps to get the best single irrep model

[On the screen: Representation analysis]
This window has an informative character. You can see here all irreps leading to non-
paramagnetic ordering. The “details” buttons provide more information about an individual
irrep and its connection to the kernel magnetic symmetry.
Press “Display representations” and look at the representation “mGM5-”.
Individual 2x2 representation matrices of each of 24 operations:

404 Jana2006 Cookbook, version October 2017

Close the window; NEXT;
[On the screen: List of kernels and epikernels]

405 Jana2006 Cookbook, version October 2017

Select e.g. magnetic group Cmcm’ and press “Show details”.

It shows symmetry operators of the selected magnetic group, transformation to the

standard setting and the relevant order parameter.

Close;
Select one of the hexagonal magnetic space groups; NEXT
This will pass all present hexagonal magnetic space groups to further testing.
[On the screen: Select Shubnikov space group]

For an individual Shubnikov space group the tool shows components of the magnetic vector
of all magnetic atoms in the structure. The button “Show details” gives a more detailed
406 Jana2006 Cookbook, version October 2017
analysis. In our case only two Shubnikov space groups – P63/mm'c' and P63' /m'm'c – allow
magnetic ordering for all Co atoms. Note that for the first of them the global magnetic
moment for is generally non-zero, which means that this arrangement allows a
ferromagnetic ordering. As ferromagnetism has not been detected for this compound, the
first model can be excluded from the analysis. However, we shall test all models to check if
such a decision can also be based on the diffraction experiment.

Select the Shubnikov space group P63/mm'c' and press “Start profile simulation”. Use "Go
to 2th/hkl" textbox and find position of the reflection 2 -1 3 . Click on the tick
corresponding to the reflection 2 -1 3:

Although the peak is clearly visible, the calculated structure factor is zero.

ESC
Press “Quit” and the select Shubnikov space groups P63'/m'm'c and investigate again the
reflection 2 -1 3:

407 Jana2006 Cookbook, version October 2017

However, in making such conclusion we have to take into account that the simulation
procedure uses randomly generated magnetic moments and that the result can be affected
by an actual random choice. This is why we recommend repeating the simulation when
some doubt exists.

For this Shubnikov space group the intensity of (2 -1 3) is non-zero in accordance with the
diffraction pattern. As this reflection is systematically absent for the nuclear structure (c
glide plane), it is clear that it originates from magnetic scattering. As apparent from the
Shubnikov space group test all magnetic moments are directed along the z-axis. This leads
to the systematic extinctions l = 2n + 1 for reflections (2h,− h,− h, l ) but only in the case that
the glide plane ”c” is combined with the time inversion operation. The fact that the peak 2 -1
3 is clearly detectable gives us another support for the Shubnikov space group P63'/m'm'c.

Select again the Shubnikov space group P63/mm'c' and press “Start graphic simulation”;
Select “Use Vesta to draw the magnetic structure”;
Press "Draw+return" or "Draw+continue";
In the case that bond between atoms are created go to “Edit→Bonds…” and delete all
automatically generated bond limits
If necessary, you can modify the length of the displayed vectors by “Edit→Vectors…”

408 Jana2006 Cookbook, version October 2017

You can see that in this case the magnetic moments, which are in the simulation chosen

409 Jana2006 Cookbook, version October 2017

 identical for the independent magnetic atoms, are not compensated in the unit cell.

Repeat this procedure for P63'/m'm'c

5. Refinement of the magnetic structure in P63/mm'c'(GM2+)

[On the screen: Select Shubnikov space group]
Select the Shubnikov space group P63/mm'c' and press the button “Continue with the
selected Shubnikov space group”;
Finish; YES to the question “Do you want to continue with the last kernel/epikernel?”;
[On the screen: Select structure name]
Use the default name "Ba5Co5_01"; OK
The test will be done under the new job name in order to keep the parent structure
unchanged. The program will open another Jana window which can be used to refine the
selected model. The Jana window with the parent structure remains active.
[On the screen: Basic window]
Right-click the icon of Refine and select the page “Various”.
Click “Fixed commands” and make sure the coordinates of all atoms are fixed:

Click “Restrictions” and make sure ADP parameters all atoms are restricted to be the same:

The “Fixed commands” and “Restrictions” are created automatically by the program when
preparing the testing structure.
OK; select the page "Basic"
Check if the “Number of cycles” is 100 and the “Damping factor” is 0.1;
“Use Marquart technique” check that the Fudge factor is 0.001;
410 Jana2006 Cookbook, version October 2017
 OK; Start refinement;
Refinement converges with Rp~3.3%, R(obs)~8.5%.

6. Refinement of the magnetic structure in P63'/m'm'c (GM4+)

Return the Jana window with the parent structure.
[On the screen: Select Shubnikov space group]
Select P63'/m'm'c and press the button “Continue with the selected Shubnikov space
group”;
[On the screen: Select structure name]
Use the default name; OK
The program will open another Jana window which can be used to refine the selected model.
The Jana window with the parent structure remains active.
[On the screen: Basic window]
Right-click the icon of Refine and select the page “Various”.
Click “Fixed commands” and make sure the coordinates of all atoms are fixed:

Click “Restrictions” and make sure ADP parameters all atoms are restricted to be the same:

OK; select the page "Basic"

Check that the “Number of cycles” is 100 and the “Damping factor” is 0.1;
“Use Marquart technique” check that the Fudge factor is 0.001;
OK; Start refinement;
Refinement converges with Rp~2.8%, R(obs)~5.6, 5.6, 3.8% for all, nuclear and magnetic
reflections, respectively.

7. Drawing of the magnetic structure

[On the screen: Basic window]
Double click on the icon “Plot structure”;
Press "Draw+return" or "Draw+continue";
Select “Objects→Structural models” and choose “Polyhedral”
Select “Edit→Bonds”; Delete bonds Ba O and Ba Cl; Apply; OK;
Select “Objects→Properties→Polyhedra”;
Choose new polyhedral style and reduce the opacity to 50:

411 Jana2006 Cookbook, version October 2017

The magnetic moments of atoms Co4 and Co5 are negligible. The atom Co6 is tetrahedrally
coordinated by O atoms while the others have octahedral coordination.

412 Jana2006 Cookbook, version October 2017

Example 12.2: Ba6Co6
Commensurate magnetic structure – modulation (propagation) vector (0,0,1/2)
Revised: 23 October 2017

Ba6Co6ClO16
Powder data measured with the ILL neutron source at 1.4 K

For more details: Mentré, O., Kauffmann, M., Ehora, G., Daviero-Minaud, S., Abraham F. &
Roussel, P. (2008). Solid State Sciences 10, 471-475.

Input files: Ba6Co6.dat (powder profile data)

Ba6Co6_nuclear.cif (nuclear structure)

1. Creating new jobname

Start Jana2006
“File → Structure → New” opens a file manager
Left pane: locate directory with input files
Right pane: double-click Ba6Co6

2. Import Wizard
Select “Magnetic: nuclear model from CIF”; NEXT
Right pane: select the input file Ba6Co6_nuclear.cif; OK

3. Define magnetic propagation vector and form factors

[On the screen: Define magnetic propagation vector and form factors]
Check “Use non-zero magnetic propagation vector” and fill its value: 0 0 1/2
Select atom type Co; check “Use as a magnetic atom”; select “Magnetic formfactor <j0>”
Co4+; NEXT; FINISH
The parent structure is created

4. Import the reflection file

[On the screen: Specify type of the file to be imported]
Select “Various CW format”; NEXT;
Check that the name of the input file is “Ba6Co6.dat” and the data format is “Free format of
2th, I, [sig(I)]”; NEXT
[On the screen: Complete/correct experimental parameters];
Fill used wavelength 2.394; NEXT; FINISH
[On the screen: Data repository]
OK; YES to “Do you want to accept made changes?”

5. Refining of profile parameters by the le Bail technique.

[On the screen: Information]
NEXT
[On the screen: Refinement of the powder profile by the le Bail algorithm]
“Show powder profile”
Although the background is often strongly varying with the diffraction angle for powder
413 Jana2006 Cookbook, version October 2017
 neutron data, in this case it is relatively smooth and it can be described without introducing
a manual background.
“Quit”; “Edit profile parameters”;
[On the screen: Powder options]
In the page “Cell” activate refinement of a and c parameters;
In the page “Profile” change the starting value of GW to 50 and activate its refinement;
In the page “Corrections” activate refinement of “shift” parameter; for background enlarge
the number of used Legendre polynomials to 10
OK; YES to save the changes
“Edit refinement commands”; In the page “Basic” change the number of cycles to 100; OK
“Run Refine”;
The refinement should converge to Rp~9.3%.
“Edit profile parameters”;
[On the screen: Powder options]
In the page “Profile” change activate refinement of GU and GV
Change the peak-shape function to pseudo-Voigt and activate refinement LY
OK; YES to save the changes
“Run Refine”
The refinement should converge to Rp~5.5%.
NEXT; YES to leave the form for le Bail refinement

6. Refinement of the scale parameter for the nuclear structure

[On the screen: Refinement of the nuclear structure by the Rietveld method]
“Edit refinement commands”; On the page “Various” select “Fixed commands”;
Fix all parameters for all atoms (wild character *); “Add”; OK; OK
“Run Refine”
The refinement should converge to Rp=9.7%, R(obs)=6.5%
NEXT; YES to leave the for for Rietveld refinement

7. Testing different irreps to get the best single irrep model

[On the screen: Information]
NEXT
[On the screen: Representation analysis]

414 Jana2006 Cookbook, version October 2017

This window has an informative character and shows all irreps leading to non-paramagnetic
ordering. The “Details” buttons provides more information about an individual irrep. There
are four highest symmetrical hexagonal Shubnikov space groups; two of them differ only in
the origin position, but this may lead to different solutions in commensurate cases.
NEXT
[On screen: List of kernels and epikernels]

Select one of the hexagonal Shubnikov groups; NEXT

This will pass all present hexagonal Shubnikov groups to further testing.
[On the screen: Select Shubnikov space group]

415 Jana2006 Cookbook, version October 2017

 Only the second and the fourth Shubnikov group allows for non-zero magnetic moments for
the Co atoms.

Select the Shubnikov group: P[c]-6c2

Press “Start profile simulation”;
Press "X exactly" and focus the region 18-40 degrees; Press “Fit Y”:

Press “Quit” and repeat this procedure for the next magnetic Shubnikov group P[c]-6c2 with
origin shifted by (0,0,1/4):

At this stage none of them can be excluded.

8. Refinement of the magnetic structure – P[c]-6c2

[On the screen: Select Shubnikov space group]
Select the Shubnikov space group P[c]-6c2 and press the button “Continue with the selected
Shubnikov space group”;
[On the screen: Select structure name]
416 Jana2006 Cookbook, version October 2017
 Use the default name; OK
[On the screen: Basic window of Jana2006]
Right-click the icon of Refine and select the page “Various”
Use the button “Restrictions” and ensure that all atoms (i.e. *) have been restricted to have
identical ADP parameters.
The restriction “restric * 12” has been set automatically when the program created the
testing structure. To see the meaning of the restriction focus it by mouse and press “Load”.
“Load” opens the restriction for editing. Make sure that "Coordinates" and "Modulations"
boxes are cleared and press “Rewrite” to close the editing mode.
OK
Select the page “Fixed commands” and (similarly like in the previous step) ensure that all
atomic positions are fixed in their current values (“fixed xyz *”).
OK
Select the page “Basic” and check if the “Number of cycles” is 100 and the “Damping factor”
is 0.1; OK; Start refinement
Refinement converges with Rp~8.8%, R(obs)~ 6.5%, 6.5%, 6.5% for all, nuclear and magnetic
reflections, respectively.

9. Drawing the resulting magnetic structure

[On the screen: Basic window]
Double click on the icon “Plot structure”
Select “Use Vesta to draw the magnetic structure”
Press "Draw+return" or "Draw+continue"
Select “Objects→Structural models” and choose “Polyhedral”
Select “Edit→Bonds”; Delete bonds Ba-O, Ba-Cl and O; Apply;OK;
Select “Objecs→Properties→Polyhedra”;
Choose new polyhedral style and reduce the opacity to 50:

417 Jana2006 Cookbook, version October 2017

10. Refinement of the magnetic structure – P[c]-6c2 origin shifted by (0,0,1/4)
Return to the Jana window with the parent structure.
[On the screen: Select Shubnikov space group]
Select P[c]-6c2+(0,0,1/4) and press the button “Continue with the selected Shubnikov space
group”;
YES to the question “Do you want to continue with the last kernel/epikernel?”;
[On the screen: Select structure name]
Use the default name; OK
[On the screen: The following transformation can bring the structure to the standard
setting]
NO to refuse the suggested origin shift
[On the screen: Basic window]
Right-click the icon of Refine and select the page “Various”.
Similarly like in the previous test ensure (under “Restrictions”) that all atoms are restricted
to have identical ADP and ensure (under “Fixed commands”) that coordinates of all atoms
are fixed.
Select the page "Basic" and check that the "Number of cycles" is 100 and the "Damping
factor" is 0.1; OK; Start refinement;
Refinement converges with Rp~8.9%, R(obs)~ 6.6%, 6.5%, 7.7% for all, nuclear and magnetic
reflections, respectively.

11. Comparing of two different results

[On screen: Basic window]
Double click on the icon “Plot structure”;
Select “Use Vesta to draw the magnetic structure”; OK
Press "Draw+return" or "Draw+continue"
In VESTA select “Objects→Structural models” and choose “Polyhedral”
Select “Edit→Bonds”; Delete bonds Ba-O, Ba-Cl and O; Apply;OK;
Select “Objecs→Properties→Polyhedra”;
Choose new polyhedral style and reduce the opacity to 50;

418 Jana2006 Cookbook, version October 2017

The models are very different. In the first model the only significant magnetic moment was
localized at the tetrahedrally coordinated Co atom, while in the second one the significantly
non-zero magnetic moment was localized at the octahedrally coordinated site.

The first structure model has better R values. Moreover, the presence of magnetic moments
at the tetrahedral site corresponds to the result in Example 12.1

419 Jana2006 Cookbook, version October 2017

420 Jana2006 Cookbook, version October 2017
Example 12.3: PrSrMnO
Commensurate magnetic structure – modulation (propagation) vector (1,0,0)
Revised: 28 November 2014

Powder data measured the ILL neutron source at 10 K

For more details: K. Knížek, J. Hejtmánek, and Z. Jirák, C. Martin, M. Hervieu, and B. Raveau,
G. André and F. Bourée, Chemistry of Materials, (2004), 1104-1110.

Input files: PrSrMnO.dat (powder profile data)

PrSrMnO_nuclear.cif (nuclear structure)

1. Creating new jobname

Start Jana2006
“File → Structure → New” opens a file manager
Left pane: locate directory with input files
Right pane: double-click PrSrMnO

2. Import Wizard
Select “Magnetic: nuclear model from CIF”; NEXT
[On the screen: Select input CIF file]
Right pane: select the input file PrSrMnO_basic.cif; OK

3. Define magnetic propagation vector and form factors

[On the screen: Define magnetic propagation vector and form factors]
Check “Use non-zero magnetic propagation vector” and fill its value: 1 0 0
For “Atom type” select “Mn”; check “Use as a magnetic atom”
For “Magnetic formfactor <j0>” select Mn3+; NEXT; FINISH
The parent structure is created

4. Import the reflection file

[On the screen: Specify type of the file to be imported]
Select “Various CW format”; NEXT
Check that the name of the input file is “PrSrMnO.dat” and the format is “ILL D1A/D2B
(Rietveld-Hewat format)”; NEXT
[On the screen: Complete/correct experimental parameters];
For “Wave length” type 1.227; NEXT; FINISH
[On the screen: Data repository]
OK; YES to “Do you want to accept made changes?”

5. Refining of profile parameters by le Bail technique.

[On the screen: Information]
NEXT;
[On the screen: Refinement of the powder profile by the le Bail algorithm]
“Show powder profile”
The background is relatively smooth and it can be described without introducing manual
background.

421 Jana2006 Cookbook, version October 2017

“Quit”;
“Edit profile parameters”;
[On the screen: Powder options]
In page “Cell” activate refinement of a, b and c parameters;
In the page “Profile” change the starting value of GW to 50 and activate its refinement;
In the page “Corrections” activate refinement of “shift” parameter and change the number
of used Legendre polynomials to 10
OK; YES to rewrite the changes
“Edit refinement command”; In the page “Basic” change the number of cycles to 20
OK; “Run Refine”;
The refinement should end with Rp~7.8%.
“Edit profile parameters”;
[On the screen: Powder options]
In the page “Profile” activate refinement of GU and GV; Change the peak-shape function to
pseudo-Voigt and activate refinement of LY; OK; YES to the question about rewriting;
“Run Refine”
The refinement should converge to Rp~5.9%.
“Show powder profile”; Select the region 2θ from 7 to 22 degs – use the button “X exactly”.

The first peak has a strong asymmetry. For neutron and synchrotron cases, such asymmetry
can be described by the correction for divergence as introduced by Finger-Cox-Jephcoat, J.
Appl. Cryst. (1994). 27, 892-900.
“Quit”
“Edit profile parameters”
[On the screen: Powder options]
In the page “Asymmetry” select “correction by divergence”

422 Jana2006 Cookbook, version October 2017

 Activate refinement of HpS/L and HmS /L keeping their default values
OK; YES to save changes
“Run Refine”;
The refinement should end with Rp~4.7%.
“Show powder profile”; Select the region 2θ from about 7 to 22 degs.

The correction considerably improved the profile

“Quit”; NEXT; YES to leave the form for le Bail refinement

6. Refinement of the scale parameter for the nuclear structure

[On the screen: Refinement of the nuclear structure by the Rietveld method]
“Edit refinement commands”; On the page “Various” select “Fixed commands”;
Fix all parameters for all atoms (“*” for “Atoms/Parameters” textbox ); “Add”; OK; OK;
The command will appear in the panel as “fixed all *”
“Run Refine”
The refinement should converge to Rp~11.6%, R(obs)~3.6%.
NEXT; YES to leave the form for Rietvel refinement

Testing different irreps to get the best single irrep model

[On the screen: Information]
NEXT
[On the screen: Representation analysis]
This window has an information character. You can see here all irreps leading to non-
paramagnetic ordering. The “details” buttons provide more information about an individual
irrep and its connection to the kernel magnetic symmetry.

423 Jana2006 Cookbook, version October 2017

NEXT
[On the screen: List of kernels and epikernels]

NEXT
[On the screen: Select Shubnikov space group]

424 Jana2006 Cookbook, version October 2017

 This form shows that only for Shubnikov space group № 2, № 4, № 6 and № 8 allows non-
zero magnetic moment for the Mn1 atom.In the following we shall test all four possibilities.
Pr1 Sr1

7. Refinement in the Shubnikov space group C[B]cce (№ 2 in the list)

[On the screen: Select Shubnikov space group]
Select the second line in the left panel; FINISH
YES to the question “Do you want to continue with the last kernel/epikernel”;
[On screen: Select structure name]
Use the default name i.e. PrSrMnO_01; OK
After selection of the actual Shubnikov group you can use either the cell and origin selection
as follows from the setting of the parent structure or you can apply transformation to get a
standard setting.
NO to the question “Do you want to apply this transformation?”.
Right click on the icon “Refine”; Select the “Basic” and change the “Number of cycles” to
100 and the “Damping factor” to 0.1;
Select the page “Various” and use the button “Restrictions”
Check that all atoms are restricted to have identical ADP parameters and not restricted
occupancies;
Restrict atoms Pr1 and Sr1, occupying the same position in the nuclear structure: identical
Coordinates ADP parameters and “keep overall sum” for occupancies; Add;

OK; “Yes+start”;
425 Jana2006 Cookbook, version October 2017
Refinement converges with Rp~10.1%, R(obs)~3.7%.
This refinement is very unstable.

Double click on the icon “Plot structure” and select Vesta to draw the structure:

Use “Tools → Powder → Profile viewer” and check fit of the calculated powder profile:

The model cannot explain the peak at (1,0,0)

“Quit”

426 Jana2006 Cookbook, version October 2017

 1Sr1

8. Refinement in the Shubnikov superspace group C[B]mcm with the origin shift (0,1/4,1/4)
(№ 4 in the list)
[On screen: Basic window]
“File → Structure → History” and select the parent structure PrSrMnO; OK;
[On screen: Basic window]
“Tools → Special tools → Representation analysis for magnetic structures”;
[On screen: Representation analysis]
NEXT; NEXT; Select the fourth Shubnikov superspace group “C[B]mcm+(0,1/4,1/4)”; FINISH
YES to the question “Do you want to continue with the last kernel/epikernel?”;
[On screen: Select structure name]
Use the default name i.e. PrSrMnO_02; OK
No to the question “Do you want to apply this transformation”.
Right click on the icon “Refine”; Select the “Basic” and change the “Number of cycles” to
100 and the “Damping factor” to 0.1;
Select the page “Various” and use the button “Restrictions”
Check that all atoms are restricted to have identical ADP parameters and not restricted
occupancies;
Restrict atoms Pr1 and Sr1, occupying the same position in the nuclear structure: identical
Coordinates ADP parameters and “keep overall sum” for occupancies; Add;
OK; OK; “Yes+start”;
Refinement converges with Rp~6.0%, R(obs)~3.2%.
Double click on the icon “Plot structure” and select Vesta to draw the structure::

Use “Tools → Powder → Profile” viewer and check fit of the calculated powder profile:

427 Jana2006 Cookbook, version October 2017

 The profile fit is very good.

9. Refinement in the Shubnikov superspace group C[B]mcm with the origin shift (0,1/4,1/4)
(№ 6 in the list)
The symbol in the table is the same as for the № 4 but both epikernels are generally
different. Compare details for both. The first one has in the original setting the symbol
A[C]mma and the second one A[C]mam. Both give the same symbol when transformed to
the standard setting.
Make the same steps and calculation as in the previous test.

Rp~8.1%, R(obs)~ 3.9%.

428 Jana2006 Cookbook, version October 2017

10. Refinement in the Shubnikov superspace group C[B]mme+(0,1/4,1/4) (№ 8 in the list)
Make the same calculation for the last Shubnikov superspace group with sin wave (Line
eight in “Select Shubnikov space group” dialogue).

Rp~11.5%, R(obs)~ 3.1%.

This refinement is very unstable.

429 Jana2006 Cookbook, version October 2017

430 Jana2006 Cookbook, version October 2017
Example 12.4: MnWO4
Incommensurate magnetic structure
Revised: 28 November 2014

Powder data measured the ILL neutron source at 10 K

For more details: L. Meddar, M. Josse, P. Deniard, C. La, G. Andre, F. Damay, V. Petříček, S.
Jobic, M.H. Whangbo, M. Maglione & C. Payen: Effect of Nonmagnetic Substituents Mg and Zn
on the Phase Competitionin the Multiferroic Antiferromagnet MnWO4
Chemistry of Materials, 21, (2009), 5203-5214.

Input files: MnWO4.dat (powder profile data)

MnWO4_nuclear.cif (nuclear structure)

1. Creating new jobname

Start Jana2006
“File → Structure → New” opens a file manager
Left pane: locate directory with input files
Right pane: double-click MnWO4

2. Import Wizard
Select “Magnetic: nuclear model from CIF”, NEXT
Right pane: select the input file MnWO4_nuclear.cif

3. Define magnetic propagation vector and form factors

[On the screen: Define magnetic propagation vector and form factors]
Check “Use non-zero magnetic propagation vector” and fill its value: -0.209 0.5 0.453
Select atom type Mn; check “Use as a magnetic atom”; select “Magnetic form factor <j0>”
Mn2+; NEXT; FINISH
The parent structure is created

4. Import the reflection file

[On the screen: Specify type of the file to be imported]
Select “Various CW format”; NEXT;
Check that the name of the input file is “MnWO4.dat” and the data format is “Free format
of 2th, I, [sig(I)]”; NEXT
[On the screen: Complete/correct experimental parameters];
Fill used wavelength 2.4226; NEXT; FINISH
[On the screen: Data repository]
OK; YES to “Do you want to accept made changes?”

5. Refining of profile parameters by the le Bail technique.

[On the screen: Information]
NEXT
[On the screen: Refinement of the powder profile by the le Bail algorithm]
“Show powder profile”
The background is relatively smooth and it can be described without introducing manual

431 Jana2006 Cookbook, version October 2017

 background.
“Quit”; “Edit profile parameters”;
[On the screen: Powder options]
In page “Cell” activate refinement of a, b, c parameters and beta angle;
Activate refinement of component q1 and q3 of the modulation vector. The q2 component
is fixed for planar monoclinic superspace groups to 0.5;
Check that the maximal satellite index is set to 1;
In the page “Profile” change the starting value of GW to 50 and activate its refinement;
In the page “Corrections” activate refinement of “shift” parameter and enlarge the number
of used Legendre polynomials to 20
OK; YES to save the changes
“Edit refinement commands”; In the page “Basic” change the number of cycles to 100 and
the damping factor to 0.5; OK
“Run Refine”;
The refinement should converge to Rp~5.2%.
“Edit profile parameters”;
[On the screen: Powder options]
In the page “Profile” change activate refinement of GU and GV
Change the peak-shape function to pseudo-Voigt and activate refinement LY
OK; YES to save the changes
“Run Refine”
The refinement should converge to Rp~3.4%.
NEXT; YES to leave the form for le Bail refinement

6. Refinement of the scale parameter for the nuclear structure

[On the screen: Refinement of the nuclear structure by the Rietveld method]
“Edit refinement commands”; On the page “Various” select “Fixed commands”;
Fix all parameters for all atoms (wild character *); “Add”; OK; OK
“Run Refine”
The refinement should converge to Rp=12.8%, R(obs)=5.5%.
NEXT; YES to leave the form for Rietveld refinement

7. Testing different irreps to get the best single irrep model

[On the screen: Information]
NEXT
[On the screen: Representation analysis]
This window has an information character. You can see here all irreps leading to non-
paramagnetic ordering. The “details” buttons provide more information about an individual
irrep and its connection to the kernel magnetic symmetry.

NEXT
[On the screen: List of kernels and epikernels]

432 Jana2006 Cookbook, version October 2017

NEXT
[On the screen: Select Shubnikov superspace group]

Notice that for both SSG’s there are non-zero amplitudes for sine and cosine waves.
However only three amplitudes are independent as can be found from “Details”:

The cosine amplitudes have generally non-zero values but they are related with the sine
amplitudes. However the coefficients are functions of y coordinate of the Mn1 atom, which
means that these restrictions are non-linear. Later we shall see how these restrictions can be
simplified.
In this case, both possibilities yield similar diffraction patterns and we cannot decide at this
stage which solution makes a better fit.

Note: The presence of a rational non-zero modulation vector component qy = 0.5 leads to a
specific problem of the origin selection. While for the space group P 2 c the c glide planes

433 Jana2006 Cookbook, version October 2017

 located at y=0 and y=1/2 are equivalent, in the SSG P 2 c .1′(α 1 2γ )00 s they are not. In
this SSG the first glide plane is connected with zero shift along the internal space while the
second one with 1/2 shift. This means that the origin selection used for the starting structure
model solved and refined from the main reflections, must be reconsidered for solution of
modulated structure. Here two models shifted by 1/2 in y must be tested. Another possibility
is to use non-standard SSG symbol P 2 c .1′(α 1 2γ )0 ss , which would generate reverse
order of c glide planes. The second method is used in this example.

8. Refinement in the first Shubnikov superspace group – P2/c.1'(a1/2g)00s

[On the screen: Select Shubnikov space group]
Select the first line in the left panel; FINISH
YES to the question “Do you want to continue with the last kernel/epikernel”;
[On screen: Select structure name]
Use the default name i.e. MnWO4_01; OK
Right click on the icon “Refine”; Select the page “Basic” and change the “Number of cycles”
to 100 and the “Damping factor” to 0.1;
Select the page “Various”. Use the button “Restrictions”. All atoms are restricted to have
identical ADP parameters;
Restrict atoms Zn1 and Mn1, occupying the same position in the nuclear structure: identical
Coordinates + Modulations + ADP parameters and “keep overall sum” for occupancies
Use button “Fixed commands” and fix all coordinates and delete or disable fixing of ADP’s
(if present).
OK; OK; “Yes+start”;
The refinement should end with Rp~4.9% and Bragg R factors R(obs)=5.3%, 2.9% and 9.1%
for all, main and satellite reflections, respectively.
Use “Tools → Powder → Profile” viewer and check fit of the calculated powder profile and
focus to the interval (38.5,47):

434 Jana2006 Cookbook, version October 2017

 Quit

9. Refinement in the second Shubnikov superspace group – P2/c.1'(a1/2g)0ss

[On the screen: Basic window]
“File → Structure → History” and select the parent structure MnWO4; OK;
[On screen: Basic window]
“Tools → Special tools → Representation analysis for magnetic structures”;
Continue with representation analysis; OK;
[On the screen: Representation analysis]
NEXT; NEXT; Select P2/c.1’(a1/2g)0ss; FINISH;
YES to the question “Do you want to continue with the last kernel/epikernel?”;
[On the screen: Select structure name]
Use the default name; OK
[On the screen: Basic window]
Right click on the icon “Refine”; Select the “Basic” and change the “Number of cycles” to
100 and the “Damping factor” to 0.1;
Select the page “Various”. Use the button “Restrictions”. All atoms are restricted to have
identical ADP parameters;
Restrict atoms Zn1 and Mn1, occupying the same position in the nuclear structure: identical
Coordinates + Modulations + ADP parameters and “keep overall sum” for occupancies
Use button “Fixed commands” and fix all coordinates and delete or disable fixing of ADP’s
(if present).
OK; OK; “Yes+start”;
The refinement should end with Rp~4.3% and Bragg R factors R(obs)=3.8%, 2.3% and 5.5%

435 Jana2006 Cookbook, version October 2017

 for all, main and satellite reflections, respectively.
Use “Tools → Powder → Profile” viewer and check fit of the calculated powder and focus to
the interval (38.5,47):

The second solution gives significantly better profile fit and also better Bragg R values.
Quit

10. Transforming the final solution to the supercentered cell.

[On the screen: Basic window]
“Tools→Transformation→Transform to the X-centered cell”;
Create new structure e.g. “MnWO4_02_X”;
YES to the question “Do you want to continue with the new structure?”;

11. Refinement in the supercentered setting

[On the screen: Basic window]
Run EditM50, go to “Symmetry”; Check the new SSG symbol; "X" stands for the
supercentered setting. For more details about centering vectors expand the list of centering
symbols at “cell centering” and re-select the symbol X
The generated supercell centering (0,1/2,0,1/2) is in accordance with the values from
“findSSG” (except that the monoclinic axis is now "b")
ESC; ESC;
Right click on the icon “Refine”; Select the “Basic” and change the “Number of cycles” to
100 and the “Damping factor” to 0.1;
Select the page “Various”; Use the button “Restrictions” and restrict all atoms to have
identical ADP parameters and not restricted occupancies; Add;
Restrict atoms Zn1 and Mn1, occupying the same position in the nuclear structure: identical
436 Jana2006 Cookbook, version October 2017
Coordinates + Modulation + ADP parameters and “keep overall sum” for occupancies;
Add; OK
Use button “Fixed commands” and fix coordinates for all atoms, i.e. *.
OK; OK; “Yes+start”;
The refinement should finish with Rp~5.0% and Bragg R factors R(obs)=4.3%, 3.0% and 6.4%
for all, main and satellite reflections, respectively.
YES to the question “Open the listing”; Check the symmetry restrictions:

Now the y component of the sine wave and x and z components of the cos wave are fixed to
zero values. This setting makes a full separation of internal and external space (before the
transformation some wave components depended on the y coordinate).

437 Jana2006 Cookbook, version October 2017

438 Jana2006 Cookbook, version October 2017
Example 12.5: HoNi
Magnetic structure refined from single crystal data – k=(0,0,1)
Revised: 23 October 2017

Single crystal data

For more details: not yet published

Input files: HoNi.hkl (powder profile data)

HoNi_nuclear.cif (nuclear structure)

1. Creating new jobname

Start Jana2006
“File → Structure → New” opens a file manager
Left pane: locate directory with input files
Right pane: double-click HoNi

2. Import Wizard
Select “Magnetic: nuclear model from CIF”, NEXT
Right pane: select the input file HoNi_nuclear.cif

3. Define magnetic propagation vector and form factors

[On the screen: Define magnetic propagation vector and form factors]
Check “Use non-zero magnetic propagation vector” and fill its value: 0 0 1
Select atom type Ho; check “Use as a magnetic atom”; select “Magnetic form factor
<j0>+c<j2>” Ho3+; NEXT; FINISH
The parent structure is created

4. Import the reflection file

[On the screen: Specify type of the file to be imported]
Select “reflection file corrected for LP and absorption”; NEXT;
Check that the name of the input file is “HoNi.hkl” and the data format is “General file on I”;
Type in the “Input format:” (3i5,2f10.2); NEXT
[On the screen: Complete/correct experimental parameters];
Fill used wavelength 0.794 Å ; NEXT; NEXT; FINISH
[On the screen: Data repository]
OK; YES to “Do you want to accept made changes?”

5. Testing different irreps to get the best single irrep model

[On the screen: Information]
NEXT
[On the screen: Representation analysis]
This window has an information character. You can see here all irreps leading to non-
paramagnetic ordering. The “details” buttons provide more information about an individual
irrep and its connection to the kernel magnetic symmetry.

439 Jana2006 Cookbook, version October 2017

 NEXT
[On the screen: List of kernels and epikernels]

Select one of the tetragonal Shubnikov space groups; NEXT

This will pass all present tetragonal Shubnikov space groups to further testing.
NEXT

6. Testing tetragonal Shubnikov space groups

[On the screen: Select Shubnikov superspace group]

440 Jana2006 Cookbook, version October 2017

 Note that only the second ShSG P[I]4nc allows for a non-zero magnetic moment.
Select it and press ‘Continue with the selected Shubnikov space group’;
[On the screen: Select structure name]
Use the default name "HoNi_01"; OK
[On the screen: Basic window of Jana2006]
Right-click the icon of Refine and select the page “Various”
Use the button “Restrictions” and ensure that all atoms (i.e. *) have been restricted to have
identical ADP parameters.
OK
Select the page “Fixed commands” and (similarly like in the previous step) ensure that all
atomic positions are fixed in their current values (“fixed xyz *”).
OK
Select the page “Basic” and check if the “Number of cycles” is 100 and the “Damping factor”
is 0.1; OK;

7. Refinement in the tetragonal Shubnikov space group – P[I]4nc

[On the screen: Basic window]
Start refinement;
Refinement converges with R(obs)~ 20.8%, 12.8%, 40.8% for all, nuclear and magnetic
reflections, respectively.
[On the screen: Basic window]
“Parameters → Extinction”
[On the screen: Extinction model]
Start “Parameters → Extinction”
Select Isotropic, Type1, Gaussian; OK; Yes to rewrite changes
Run refinement
Refinement converges with R(obs)~ 17.4%, 8.2%, 40.8% for all, nuclear and magnetic
reflections, respectively.
Right click on the icon “Refine”;

441 Jana2006 Cookbook, version October 2017

 Select the page “Various” and use the button “Restrictions”;
Disable the command “restrict * 12” which keeps ADP's of all atoms identical. OK;
Use the button “Fixed commands” and disable the command “fixed xyz *” which fixes
coordinates of all atoms. OK; OK; Yes+start;
Refinement converges with R(obs)~ 16.7%, 7.7%, 39.4% for all, nuclear and magnetic
reflections, respectively.
[On the screen: Basic window]
Double click on the icon “Plot structure”; Use the Vesta program to draw the structure;
Press "Draw+return" or "Draw+continue";
In Vesta enlarge the boundaries in z direction:

8. Testing orthorhombic Shubnikov space group C[B]mc21

[On the screen: Basic window of the parent structure, with the representation analysis]
“Back → Select the space group C[B]mc21 → Next;

Select the only space group and press ‘Continue with the selected Shubnikov space group’;
[On the screen: Select structure name]
Use the default name "HoNi_02"; OK
442 Jana2006 Cookbook, version October 2017
NO to the question “Do you want to apply this transformation?”;
[On the screen: Basic window]
Start refinement;
Refinement converges with R(obs)~ 12.8%, 12.3%, 14.3% for all, nuclear and magnetic
reflections, respectively.
[On the screen: Basic window]
“Parameters → Extinction”
[On the screen: Extinction model]
Start “Parameters → Extinction”
Select Isotropic, Type1, Gaussian; OK; Yes to rewrite changes
[On the screen: Basic window]
Run refinement
Refinement converges with R(obs)~ 9.6%, 7.8%, 14.6% for all, nuclear and magnetic
reflections, respectively.
Right click on the icon “Refine”;
Select the page “Various” and use the button “Restrictions”;
Disable the command “restrict * 12” which keeps ADP's of all atoms identical. OK;
Use the button “Fixed” and disable the command “fixed xyz *” which fixes coordinates of all
atoms. OK; OK; Yes+start;
Refinement converges with R(obs)~ 8.4%, 6.7%, 13.0% for all, nuclear and magnetic
reflections, respectively.
Start “Parameters → Twin fractions”;
[On the screen: Edit twin fraction]
Selection of the orthorhombic Shubnikov space group represents a lowering symmetry from
the original one (parent symmetry. This can create different magnetic domains related by
the lost symmetry operations. The lost symmetry operation has been automatically created
by the program as a twinning operation. The default value of the relative fraction was 0.5
which means 1:1 ratio. In the final stage this value can be refined.
Press the button “Show twinning matrix”.
[On the screen: Twinning matrix #2]
As expected the twinning operation is the four-fold axis, which was present in the tetragonal
description but it is missing in the orthorhombic description.
OK;
[On the screen: Edit twin fraction]
Change the volume fraction to zero and activate the refinement key for the volume fraction.
OK; Yes to the question “Do you want to rewrite changed files?”.
[On the screen: Basic window]
Run refinement
Refinement converges with R(obs)~ 8.3%, 6.7%, 12.6% for all, nuclear and magnetic
reflections, respectively.
[On the screen: Basic window]
Start “Parameters → Twin fractions”;
[On the screen: Edit twin fraction]
Check that the volume fraction return to the value close to 0.41.
[On the screen: Basic window]
Double click on the icon “Plot structure”; Use the Vesta program to draw the structure;
Press "Draw+return" or "Draw+continue";
443 Jana2006 Cookbook, version October 2017
 In Vesta enlarge the boundaries in z direction (like in the previous plot):

9. Testing orthorhombic Shubnikov space group P[I]mn21

[On the screen: Basic window]
Repeat the procedure from the previous point but now for the Shubnikov space group
P[I]mn21.
Refinement converges with R(obs)~ 8.8%, 7.4%, 12.4% for all, nuclear and magnetic
reflections, respectively.

444 Jana2006 Cookbook, version October 2017

Example 12.6: ISODISTORT
Practical introduction to Jana2006-ISODISTORTcommunication
Revised: 23 October 2017

Input files: Ba5Co5.m40, Ba5Co5.m41, Ba5Co5.m50, Ba5Co5.m90, Ba5Co5.m95

1. Creating CIF file from the parent structure

[On the screen: Basic window]
Start Jana2006
Start “File → Structure → Open” and open the structure Ba5Co5;
“Tools for magnetic structures → Run ISODISTORT wizard”;

2. Start ISODISTORT home page

[On the screen: ISODISTORT=>Jana2006]
Press “Run ISODISTORT”;
Your Internet default browser will open the page:
http://stokes.byu.edu/iso/isodistort.php

Browse for the CIF file “Ba5Co5.cif”; OK

[On the screen: ISODISTORT search]

445 Jana2006 Cookbook, version October 2017
Define distortions to be considered:
In our case we are interested just in magnetic ordering;

Press Change;

446 Jana2006 Cookbook, version October 2017

In the box for "Method 2" select GM k16(0,0,0), which means k=(0,0,0):

Press OK;
Select irreducible representation
[On the screen: ISODISTORT irreducible representation ]

Comparing with the "Representation analysis" list of Jana2006, only irreducible

representations allowing non-zero magnetic moments are listed. In the next step we will
demonstrate the two dimensional irreducible representation mGM5+.
Select the representation mGM5+ and press OK;

447 Jana2006 Cookbook, version October 2017

Representation analysis of Jana2006 (for comparison):

[On the screen: ISODISTORT: order parameter direction]

Here we can see the kernel and the set of epikernels (isotropy subgroups) associated with
mGM5+.

448 Jana2006 Cookbook, version October 2017

List of kernels and epikernels in Jana2006 (for comparison). The order parameters may differ
depending on the setting of the basis.

Return one step back in your Internet browser and select the representation mGM2+; OK;
[On the screen: ISODISTORT: order parameter direction]

Select the listed ODP and press OK

[On the screen: ISODISTORT: distortion]

Select a small starting values (0.1) for magnetic moments of Co4, Co4 and Co6;
Select “CIF file” and press OK;
Save it as a Text Document with name GM2+ into the working directory.

449 Jana2006 Cookbook, version October 2017

Select “View distortion” and press OK;

Make the steps related with usage of Java.

If Java offers an update please do it, otherwise the applet may be blocked. In some cases the
Java security level needs to be adjusted in Java Control Panel (and the browser restarted and
the work repeated ...).

450 Jana2006 Cookbook, version October 2017

 Using the sliders, magnetic moments of Co atoms can be changed to see mutual
arrangement of magnetic moments of symmetry related atoms.

Return back to the page ISODISTORT irreducible representation and select now mGM4+

Repeat the work as for GM2+ and save the CIF file as a Text Document with name GM4+
into the working directory.
Return to Jana2006

3. Refining the model mGM2+

[On the screen: ISODISTORT => Jana2006]
Select the magnetic space group P63/mm’c’ irrep: mGM2+
Press button “Continue with the selected isotropic subgroup”;
Select structure name e.g. Ba5Co5_01;
[On the screen: Basic window]
Run Refine;
Refinement converges with Rp~2.9%, R(obs)~8.1%.

451 Jana2006 Cookbook, version October 2017

4. Refining the model mGM4+
[On the screen: Basic window]

[On the screen: ISODISTORT => Jana2006]

Select the magnetic space group P63’/m’m’c irrep: mGM4+
Press button “Continue with the selected isotropic subgroup”;
Select structure name e.g. Ba5Co5_02;
[On the screen: Basic window]
Run Refine;
Refinement converges with Rp~2.6%, R(obs)~5.4%.
The results are identical with those from example 12.1.

452 Jana2006 Cookbook, version October 2017

Example 12.7: DyMn6Ge6
Incommensurate magnetic structure – k=(0,0,0.1651); k=(0,0,0)
Revised: 23 October 2017

DyMn6Ge6
Neutron powder data measured at 11K on DMC at PSI (Switzerland)
Available at the Bilbao crystallographic server, database magndata, #1.1.10

References:
J. Rodríguez-Carvajal & F. Bourée, EPJ Web of Conferences 22, 00010 (2012)
P. Schobinger-Papamantellos et al, Journal of Alloys and Compounds, 203 (1994) 243

Input files: dymn6ge6.dat (powder profile data)

dymn6ge6_nuclear.cif (nuclear structure)

1. Creating new jobname

Start Jana2006
“File → Structure → New” opens a file manager
Left pane: locate directory with input files
Right pane: double-click dymn6ge6

2. Import Wizard
Select “Magnetic parent structure: nuclear model from CIF”; NEXT
Right pane: select the input file dymn6ge6_nuclear.cif; OK

3. Define magnetic propagation vector and form factors

[On the screen: Define magnetic propagation vector and form factors]
Choose the “Number of non-zero magnetic propagation vectors” to ‘1’ and fill its value: 0 0
0.1651
Select atom type Dy; check “Use as a magnetic atom”; select “Magnetic formfactor <j0>”
Dy3+;
Select atom type Mn; check “Use as a magnetic atom”; select “Magnetic formfactor <j0>”
Mn3+;
NEXT; FINISH
The parent structure is created

4. Import the reflection file

[On the screen: Specify type of the file to be imported]
Select “Various CW format”; NEXT;
Check that the name of the input file is “dymn6Ge6.dat” and the format is “PSI format”;
Select “Debye-Scherrer method”
NEXT
[On the screen: Complete/correct experimental parameters];
Fill used wavelength 1.7037; Temperature 11; NEXT; FINISH
[On the screen: Data repository]
OK; YES to “Do you want to accept made changes?”
453 Jana2006 Cookbook, version October 2017
5. Refining of profile parameters by the le Bail technique.
[On the screen: Information]
NEXT
[On the screen: Refinement of the powder profile by the le Bail algorithm]
“Show powder profile”
The background is relatively smooth and it can be described without introducing manual
background.
“Quit”;
“Edit profile parameters”;
[On the screen: Powder options]
In page “Cell” activate refinement of a and c parameters and the third component of he
modulation vector – q3;
In the page “Profile” change the starting value of GW to 50 and activate its refinement;
In the page “Corrections” activate refinement of “shift” parameter and change the number
of used Legendre polynomials to 10
OK; YES to save the changes
“Edit refinement commands”; In the page “Basic” change the number of cycles to 20; OK
“Run Refine”;
The refinement should converge to Rp~4.95%.
“Edit profile parameters”;
[On the screen: Powder options]
In the page “Profile” activate refinement of GU and GV
Change the peak-shape function to pseudo-Voigt and activate refinement LX
OK; YES to save the changes
“Run Refine”
The refinement should converge to Rp~2.98%.
“Show powder profile”
Select the region 2θ from 30 to 45 degs – use the button “X exactly”. A couple of observed
peaks may correspond to some impurity. Also at about 2θ= 5, an anomaly in the background
is observed. No peaks are possible there.
“Quit”;
“Edit profile parameters”;
[On the screen: Powder options]
In page “Corrections”, press button “Define excluded regions” : from 0 to 7; “Add”,
“OK”,“OK”;
“Run Refine”;
The refinement should end with Rp~2.90%.
NEXT; YES to leave the form for le Bail refinement

6. Refinement of the scale parameter for the nuclear structure

[On the screen: Refinement of the nuclear structure by the Rietveld method]
“Edit refinement commands”; On the page “Various” select “Fixed commands”;
In the case that the fixed command of form is “fixed all *” is already in the panel, do
nothing. If not:
Fix all parameters for all atoms (“*” for “Atoms/Parameters” textbox); “Add”; OK; OK
“Run Refine”
The refinement should converge to Rp=11.6%, R(obs)=7.5%
NEXT; YES to leave the for for Rietveld refinement
454 Jana2006 Cookbook, version October 2017
7. Testing different irreps to get the best single irrep model
[On the screen: Information]
NEXT
[On the screen: Representation analysis]

This window has an information character. You can see here all irreps leading to magnetic
ordering. The “details” buttons provide more information about an individual irrep and its
connection to the kernel magnetic symmetry.
NEXT

455 Jana2006 Cookbook, version October 2017

[On screen: List of kernels and epikernels]

Select one of the hexagonal Shubnikov superspace groups based on average space group
P6/mmm; NEXT;
This will pass all present hexagonal Shubnikov superspace groups of type P6/mmm to
further testing. We assume that both magnetic atoms should have non-zero magnetic
moment.
[On the screen: Select Shubnikov space group]

Only the second Shubnikov superspace group allows for non-zero magnetic moments for Dy
atom.

Select the second Shubnikov superspace group: P6/mmm1’(00g)00sss and press “Start
profile simulation”;
456 Jana2006 Cookbook, version October 2017
Press "X exactly" and focus the region 8-16 degrees; Press “Fit Y”:

Two dominant magnetic peaks in this region have zero intensity. This means that the tested
model does not describe the magnetic ordering correctly. However, in making such
conclusion we have to take into account that the simulation procedure uses randomly
generated magnetic moments and that the result can be affected by an actual random
choice. This is why we recommend repeating the simulation when some doubt exists.
“Quit”; “Back” and return to the “List of kernels and epikernels” and select one of
hexagonal superspace groups based on P622.
While the first one P622.1’(00g)t00s restricts the magnetic moments of the Dy atom to zero,
the second one P622.1’(00g)-h00s allows non-zero magnetic moments for both magnetic
atoms. The simulation profiles demonstrate that the intensity of both magnetic peaks in the
interval (8-16) degs are generally non-zero:

457 Jana2006 Cookbook, version October 2017

458 Jana2006 Cookbook, version October 2017
8. Refinement of the magnetic structure in – P622.1'(00g)-h00s
[On the screen: Select Shubnikov space group]
Select the Shubnikov superspace group P622.1'(00g)-h00s
Press ‘Continue with the selected magnetic superspace group’
[On the screen: Select structure name]
Use the default name i.e. dymn6ge6_01 or any other; OK
[On the screen: Basic window of Jana2006]
Right-click the icon of Refine and select the page “Various”
Use the button “Restrictions” and ensure that all atoms (i.e. *) are restricted to have
identical ADP parameters.
The restriction “restric * 12” has been set automatically when the program created the
testing structure. To see the meaning of the restriction focus it by mouse and press “Load”.
“Load” opens the restriction for editing. Make sure that "Coordinates" and "Modulations"
boxes are cleared and press “Rewrite” to close the editing mode.
OK
Select the page “Fixed commands” and (similarly like in the previous step) ensure that all
atomic positions are fixed in their current values (“fixed xyz *”).
OK
Select the page “Basic” and check if the “Number of cycles” is set to 100 and the “Damping
factor” to 0.1; OK; Yes+start;
Refinement converges with Rp~4.7%, R(obs)~ 4.9%, 4.5%, 5.8% for all, nuclear and magnetic
reflections, respectively.
Use “Profile viewer” and check the fit of the calculated powder profile
Observe that some main reflections like (1 0 1 0) around 25 degrees or (2 -1 1 0) at about 40
degrees are poorly fit. In order to exclude some additional nuclear effects we can allow the
refinement of atomic positions:
Right-click the icon of Refine and select the page “Various”. Use the button “fixed
commands” and disable the command “fixed xyz”. Run refine again.
Now the atom coordinates will also be refined. The refinement should end with Rp~4.7% and
Bragg R factors R(obs)=4.8%, 4.4% and 5.7% for all, main and satellite reflections,
respectively. (“View of Refine”). The improvement is minimal. Some of the main reflections
remain poorly fit systematically with too low intensity. This fact can lead to the conclusion
that an additional magnetic ordering with k=0 is present.
9. Refinement in the subgroups allowing magnetic wave (0,0,0)
[On the screen: Basic Jana window]
We try then the possibility that magnetic scattering occurs also at main reflections, i.e. an
additional magnetic ordering with k=(0 0 0) is taking place. In the language of the symmetry
group of the phase this implies to lower it, dropping the operation {1’|0001/2} associated
with the existence of a single primary incommensurate wave vector.

There are several maximal subgroups without this operation. Then can be generated by the
Jana tool “Tools→Transformations→Go to subgroup structure”. As generators for the
working group until now can be used:
{6+|0 0 0 5/6}: x1-x2, x1, x3, x4+5/6
{2_xy| 0 0 0 0}: x2, x1, -x3, -x4
{1’|0001/2} : x1, x2, x3 , x4+1/2
459 Jana2006 Cookbook, version October 2017
The possible maximal subgroups will then be generated by:

P622(00g)-h00
{6+|0 0 0 5/6}: x1-x2, x1, x3, x4+5/6, m
{2_xy| 0 0 0 0}: x2, x1, -x3, -x4, m

P6’22’(00g)t00
{6+’|0 0 0 1/3}: x1-x2, x1, x3, x4+1/3, -m
{2_xy| 0 0 0 0}: x2, x1, -x3, -x4, m

P62’2’(00g)-h00
{6+|0 0 0 5/6}: x1-x2, x1, x3, x4+5/6, m
{2’_xy| 0 0 0 1/2}: x2, x1, -x3, -x4+1/2, -m

P6’2’2(00g)t00
{6+’|0 0 0 1/3}: x1-x2, x1, x3, x4+1/3, -m
{2’_xy| 0 0 0 1/2}: x2, x1, -x3, -x4+1/2, -m

Make sure that your actual structure is the final refined structure from the previous step.

“Tools → Transformation → Go to subgroup structure”;

Press the “Refresh” button and select the generators:

x1-x2 x1 x3 x4+5/6 m
x2 x1 -x3 -x4 m

460 Jana2006 Cookbook, version October 2017

Press the button “Complete subgroup”;

NEXT; NEXT
[On the screen: Select one representative of the 2nd coset]
Select the first operation.
Generally, all selections would give the same result.
NEXT;
[On the screen: Specify the output structure]
Select “dymn6ge6_01_subgroup_1” or any other
NEXT;FINISH;”Yes” to the question “Do you want to continue with the new structure?”
[On the screen: Basic window]
Run “Edit Atoms”; Select atoms Dy1 and Mn1; Press the button “Action”
Select “Reset modulation/magnetic parameters”
Deselect “Magnetic parameters k<>0”
Choose “Reset parameters randomly in the interval (-pmax,pmax)
OK;OK;”Yes” to the question “Do you want to rewrite changed files?”
Run “Refine”;
Repeat the procedure for all subgroups listed above
Use the “File→ Structure → History” to return to the final refined structure in
P622.1'(00g)-h00s.

P622(00g)-h00
Rp=4.54%, R(obs)=4.75%,4.35%,5.47% for all, nuclear and magnetic reflections, respectively.

P6’22’(00g)t00
Rp=4.42%, R(obs)=4.46%,3.96%,5.38% for all, nuclear and magnetic reflections, respectively.
461 Jana2006 Cookbook, version October 2017
P62’2’(00g)-h00
Rp=3.84%, R(obs)=3.11%,2.44%,4.35% for all, nuclear and magnetic reflections, respectively.

P6’2’2(00g)t00
Rp=4.45%, R(obs)=3.91%,4.35%,5.40% for all, nuclear and magnetic reflections, respectively.

One of the symmetries clearly fits the data better. Observe that the strong misfit of some of
the main reflections has disappeared.

462 Jana2006 Cookbook, version October 2017

 The final common Uiso is negative but its absolute value is very small.
Right-click the icon of Refine and select the page “Various”
Use the button “Fixed commands”
Select “Set individual’; File the edit box by the text “Uiso[*]”; Define value “Set to” zero
Add; OK; OK; Press Button “Yes+start”
Rp=3.88%, R(obs)=3.37%,2.81%,4.42% for all, nuclear and magnetic reflections, respectively.
Run “EditM50”; Go to the page “Magnetic parameters”; Change the magnetic form factor
for Dy to <j0>+c<j2> and Dy3+
OK; ”Yes” to the question “Do you want to rewrite M50 file?”
Run “Refine”
Rp=3.83%, R(obs)=2.93%,2.43%,3.82% for all, nuclear and magnetic reflections, respectively.

10. Visualize the refined model with MVISUALIZE (Jmol)

[On the screen: Basic window]
“File → CIF utilities → Make mCIF file for MVisualize”;
[On the screen: Define name for the out mcif file]
Define here the name of a mcif file with the file extension “mcif”
OK
[On the screen: INFORMATION]
OK
The Jana will now start your default internet browser on the home page of Bilbao
Crystallographic server. Select there the utility MVISUALIZE and follow their instructions.

463 Jana2006 Cookbook, version October 2017

11. Visualize the refined model with VESTA
[On the screen: Basic window]
Double click on the icon “Plot structure”
Select “Draw approximant structure”; x from 0 to 1; y form 0 to 1; z from 0 to 6 (c*
component of the propagation vector is close to 1/6)
Select “Use Vesta to draw the magnetic structure”
Press "Draw+return" or "Draw+continue"
Select “Edit→Vectors” and modify “Scale factor for modulus” to 0.3

OK
Press button “Properties…”; Go the page “Atoms” and modify “Radius” for all atoms

464 Jana2006 Cookbook, version October 2017

465 Jana2006 Cookbook, version October 2017
12. Transformation to the standard setting as used in MAGNDATA
[On the screen: Basic window]
“Tools → Connect through internet to → Bilbao Crystallographic Server”
Select “MAGNDATA”; Enter the label of the structure “1.1.10”

The result as derived step by step from the representation theory is expressed in a different
chiral domain than used in MAGNDATA. Therefore, we should transfer it to the equivalent
chiral domain.
[On the screen: Basic window]
“Tools → Transformation → Change enantiomorph”
Create a new structure with name e.g. “dymn6ge6_01_subgroup_3_chiral”
OK; Yes to continue with the new structure
Run EdiM50; Go the page “Symmetry”
Now the superspace group is the same as used in MAGNDATA however there is a non-zero
origin shift in x4 equal 1/6.
ESC
“Tools → Transformation → Origin shift”
Use shift to the point 0 0 0 -1/6
OK
The last point that has to be checked for a full comparison is if we are using in the same
symmetry positions for the magnetic atoms Dy1 and Mn1. The positions of these atoms
from 1.1.10 are:
466 Jana2006 Cookbook, version October 2017
Dy1 (0,0,0) and Mn1 (0.5,0,0.251).
[On the screen: Basic window]
Run “Edit atoms”; Double click on Dy1 atom; Go to the page “Edit”
The coordinates are (1,1,1) and this means that we can just rewrite them to (0,0,0) as such
transformation does not change the modulation functions.
Move to the atom Mn1
Its coordinates are (0.5,1,0.749). From the list of symmetry operations we can find that the
symmetry operation: x1-x2,-x2,-x3,-x4+1/2 combined with the lattice translation (1,1,1,0)
makes what is needed.
ESC;ESC
[On the screen: Select atoms to be used]
Select the atom Mn1; Press button “Action”; Select “Transform selected atoms”
Use the button “Fill by a symmetry operation”; Select x1-x2 -x2 -x3 -x4+1/2 –m; Modify the
translation vector to 1 1 1 1/2
OK; OK; Yes to rewrite files
[On the screen: Basic window]
Run “Refine” to check that all transformation did not change the solution
[On the screen: Basic window]
“Edit/View → View of Refine”; Press “Go to” and select “Changes overview”

M0x M0y M0z |M|

MAGNDATA 0.0 0.0 3.93 3.93
Dy1
Jana 0.0 0.0 3.88(9) 3.88(9)
MAGNDATA 0.0 0.0 -1.15 -1.15
Mn1
Jana 0.0 0.0 -0.94(7) 0.94(7)

467 Jana2006 Cookbook, version October 2017

 cos sin
Mx My Mz Mx My Mz
MAGNDATA 6.20 0.0 0.0 3.58 7.15 0.0
Dy1
Jana 6.02 0.0 0.0 3.48(4) 6.95 0.0
MAGNDATA -1.88 -0.21 0.0 -0.85 -2.05 0.0
Mn1
Jana -1.44(10) -0.62(9) 0.0 -1.29(8) -2.22(8) 0.0

13. Use of the option of polar coordinates for the moments to constraint further the Mn spin
to have a circular modulation.
[On the screen: Basic window]
“File → Structure → Save as”; Use e.g. “dymn6ge6_01_subgroup_3_final”
OK; No to the question “Do you want to continue with the new structure?”
[On the screen: Basic window]
“Parameters → Magnetic moments in polar coordinates”;
Select both atoms
OK; Yes to rewrite changed files
Run “Edit atoms”
[On the screen: Select atoms to be used]
Double click on the atom Dy1; Go the the “Edit” page;
The windows for the magnetic moment parameters have changed to polar coordinates, but
the control of the parameters to be refined does not change automatically, and must be
done manually: for the Mo parameters of Dy1, fix the two polar angles, and leave to refine
the modulus (the first one). For the modulation parameters, modify the refinement controls
if necessary. Only Mrsin1 parameters should be free, the rest are forced by symmetry
namely Mrcos1=Mrsin1
For Mn1, the modulation parameters except the theta angle are symmetry free. In order to
force a circular modulation, the constraints must be used: Mrcos1=Mrsin1, Mpcos1=Mpsin1-
90. These equations are only roughly satisfied by the previous refined model, as shown after
the transformation.

Right click on the “Refine” icon; Select the page Various; Click on the button “Equations”;
Define the equations mentioned above:

Mp0[Dy1]=0
Mt0[Dy1]=0
Mrcos1[Dy1]=Mrsin1[Dy1]
Mpsin1[Dy1]=90
Mpcos1[Dy1]=0
Mtsin1[Dy1]=90
Mtcos1[Dy1]=90
Mp0[Mn1]=0
Mt0[Mn1]=180
Mrcos1[Mn1]=Mrsin1[Mn1]
Mpcos1[Mn1]= Mpsin1[Mn1]-90
Mtsin1[Mn1]=90

468 Jana2006 Cookbook, version October 2017

 Mtcos1[Mn1]=90

OK;OK; Yes+start
Rp=3.83%, R(obs)=2.93%,2.43%,3.82% for all, nuclear and magnetic reflections, respectively.
Which means that it is not significantly worse than the full refinement – see point 9.

M0x M0y M0z |M|

MAGNDATA 0.0 0.0 3.93 3.93
Dy1
Jana 0.0 0.0 3.97(10) 3.97(10)
MAGNDATA 0.0 0.0 -1.15 -1.15
Mn1
Jana 0.0 0.0 -1.06(7) -1.06(7)

After transforming from polar to crystallographic coordinates

cos sin
Mx My Mz Mx My Mz
MAGNDATA 6.20 0.0 0.0 3.58 7.15 0.0
Dy1
Jana 6.12 0.0 0.0 3.53(4) 7.07 0.0
MAGNDATA -1.88 -0.21 0.0 -0.85 -2.05 0.0
Mn1
Jana -1.76(10) -0.19(9) 0.0 -0.79(9) -1.92(8) 0.0

General note: Two different strategies for handling refinement keys in Jana2006
From Petříček, V., Dušek, M. & Palatinus, L. (2014) Z.Kristallogr. 229, 345-352.

“All refinable parameters are divided into two groups: automatic and user-controlled ones. The
automatic parameters are those which can be usually refined without numerical problems: scale
factors and extinction parameters; parameters describing position, ADP and their modulation
for individual atoms as well as for molecules; background parameters of the powder profile;
magnetic parameters. The user-controlled parameters are twin/phase volume fractions, atomic
or molecular site occupancies, parameters of crenel modulation functions, parameters of
multipole refinement and powder profile parameters for individual Bragg reflections.
The automatic parameters are activated by default at the beginning of the refinement process
while the user-controlled ones must be always explicitly activated by the user. The program
always fixes the parameters which cannot be refined due to symmetry restrictions. In order to
fix other automatic parameters the user is provided with a range of special fixing commands.
This way is very useful for most cases because it minimizes mistakes in setting refinement flags.
However, the program also allows switching to a completely manual mode, where all
parameters are to be activated explicitly by the user, analogically to other refinement programs
such as FullProf [29].”

Handling of symmetry restrictions of magnetic moments in polar coordinates would be easier

for most of users if the non-automatic procedure is used:

469 Jana2006 Cookbook, version October 2017

470 Jana2006 Cookbook, version October 2017
Example 12.8: Cr2WO6
Commensurate magnetic structure – k=(0,0,0)
Revised: 23 October 2017

Neutron powder data measured at 4K on HB2A at HFIR, λ=2.4097Å

Available at the Bilbao crystallographic server, database magndata, #0.75

References:
Zhu et al., PRL 113, 076406 (2014)
PHYSICAL REVIEW B 92, 094419 (2015)

Input files: cr2wo6.dat (powder profile data)

cr2wo6_nuclear.cif (nuclear structure)

Parent space group:

P42/mnm (N. 136)

Propagation vector:
k= (0,0,0)

Magnetic site:
Cr 4e (0,0,z)

1. Creating new jobname

Start Jana2006
“File → Structure → New” opens a file manager
Left pane: locate directory with input files
Right pane: double-click cr2wo6

2. Import Wizard
Select “Magnetic parent structure: nuclear model from CIF”; NEXT
Right pane: select the input file cr2wo6_nuclear.cif; OK

3. Define magnetic propagation vector and form factors

[On the screen: Define magnetic propagation vector and form factors]
Select “Number of non-zero magnetic propagation vectors” to 1
Select atom type Cr; check “Use as a magnetic atom”; select “Magnetic formfactor <j0>”
Cr3+;
471 Jana2006 Cookbook, version October 2017
 NEXT; FINISH
The parent structure is created

4. Import the reflection file

[On the screen: Specify type of the file to be imported]
Select “Various CW format”; NEXT;
Check that the name of the input file is “cr2wo6.dat” and the format is “free format of 2th,
I, [sig(I)]”;
NEXT
[On the screen: Complete/correct experimental parameters];
Fill used wavelength 2.4097; Temperature 4; NEXT; FINISH
[On the screen: Data repository]
OK; YES to “Do you want to accept made changes?”

5. Refining of profile parameters by the le Bail technique.

[On the screen: Information]
NEXT
[On the screen: Refinement of the powder profile by the le Bail algorithm]
“Show powder profile”
The background is relatively smooth and it can be described without introducing manual
background.
“Quit”;
“Edit profile parameters”;
[On the screen: Powder options]
In page “Cell” activate refinement of a and c parameters;
In the page “Profile” change the starting value of GW to 50 and activate its refinement;
In the page “Corrections” activate refinement of “shift” parameter and change the number
of used Legendre polynomials to 8
OK; YES to save the changes
“Edit refinement commands”; In the page “Basic” change the number of cycles to 20; OK
“Run Refine”;
The refinement should converge to Rp~17.5%.
“Edit profile parameters”;
[On the screen: Powder options]
In the page “Profile” activate refinement of GU and GV
Change the peak-shape function to pseudo-Voigt and activate refinement LX
OK; YES to save the changes
“Run Refine”
The refinement should converge to Rp~7.2%.
“Show powder profile”
Select the region 2θ from 10 to 20 degs – use the button “X exactly”. The peak (0 0 1) has
some asymmetry. “Quit”;
“Edit profile parameters”;
[On the screen: Powder options]
In page “Asymmetry/Difractometer”
Select “Berar-Baldinozzi correction”; Activate refinement of asym1, asym2, asym3, and
asym4; “OK”;
“Run Refine”;

472 Jana2006 Cookbook, version October 2017

 The refinement should end with Rp~6.5%.
NEXT; YES to leave the form for le Bail refinement

6. Refinement of the scale parameter for the nuclear structure

[On the screen: Refinement of the nuclear structure by the Rietveld method]
“Edit refinement commands”; On the page “Various” select “Fixed commands”;
In the case that the fixed command of form is “fixed all *” is already in the panel, do
nothing. If not:
Fix all parameters for all atoms (“*” for “Atoms/Parameters” textbox); “Add”; OK; OK
“Run Refine”
The refinement should converge to Rp=14.2%, R(obs)=6.8%
NEXT; YES to leave the for for Rietveld refinement

7. Testing different irreps to get the best single irrep model

[On the screen: Information]
NEXT
[On the screen: Representation analysis]

This window has an information character. You can see here all irreps leading to magnetic
ordering. The “details” buttons provide more information about an individual irrep and its
connection to the kernel magnetic symmetry.
NEXT

[On screen: List of kernels and epikernels]

473 Jana2006 Cookbook, version October 2017

Select one of the tetragonal Shubnikov space groups; NEXT;
This will pass all present tetragonal Shubnikov space groups of type P42/mnm to further
testing. We assume that Cr magnetic atoms should have non-zero magnetic moment.
[On the screen: Select Shubnikov space group]

Select the second Shubnikov superspace group: P42’/mnm’ and press “Start profile
simulation”;
Press "X exactly" and focus the region 12-33 degrees; Press “Fit Y”:

474 Jana2006 Cookbook, version October 2017

 While for the strong magnetic reflection (0,0,1)the model gives zero, for the reflection
(0,0,1) which seems to be very weak the model gives non-zero intensity. This can be
reproduced for all tetragonal Shubnikov groups allowing non-zero magnetic moments for Cr
atom – P42/mn’m’, P42/m’n’m’ and P42’/m’nm’. As a next step we should try orthorhombic
space groups.
Quit;

8. Testing orthorhombic subgroups

[On the screen: Basic window of the parent structure, with the representation analysis]
“Back → Select one of the orthorhombic C centered Shubnikov space groups;
The symbol of Shubnikov space group is related to the lattice defined by axes as expressed in
the second column. However the final transformation to the standard setting is offered only
during the last step of the representation analysis to keep relationships of magnetic
moments to the original call of the parent structure.
→ Next;

475 Jana2006 Cookbook, version October 2017

The first listed Shubnikov space group allows ferromagnetic ordering as seen from non-zero
the global magnetic moment. This global magnetic moment can be suppressed only if
additional condition to relate the magnetic moments of two independent Cr positions is
applied. This makes this model less probable as no tendency to make ferromagnetic ordering
was detected.

Switch to the second Shubnikov space group

The second Shubnikov space group has the global magnetic moment zero.

Select the first Shubnikov superspace group: Cm’mm’ and press “Start profile simulation”;
Press "X exactly" and focus the region 12-33 degrees; Press “Fit Y”;

Both reflections (1,0,0) and (0,0,1) have generally non-zero intensity and this means that we
cannot reject this space groups only from violation of systematically absent reflection as for
the tetragonal case. This is also true for Cmm’m.

Select the Shubnikov space group Cm’mm’

Press ‘continue with the selected Shubnikov space group’
[On the screen: Select structure name]
Use the default name i.e. cr2wo6_01 or any other; OK
YES to the question “Do you want to apply this transformation?”
This will transform the structure to the standard setting. Jana2006 can generally work at
non-standard setting.
[On the screen: Basic window of Jana2006]
Right-click the icon of Refine and select the page “Various”

476 Jana2006 Cookbook, version October 2017

 Use the button “Restrictions” and ensure that all atoms (i.e. *) are restricted to have
identical ADP parameters.
The restriction “restric * 12” has been set automatically when the program created the
testing structure. To see the meaning of the restriction focus it by mouse and press “Load”.
“Load” opens the restriction for editing. Make sure that "Coordinates" box are cleared and
press “Rewrite” to close the editing mode.
OK
Select “Fixed commands” and (similarly like in the previous step) ensure that all atomic
positions are fixed in their current values (“fixed xyz *”).
OK
Select the page “Equation”;
Introduce the equation b=a valid for diagonals in tetragonal lattice;
Select the page “Basic” and check that the “Number of cycles” is set to 100 and the
“Damping factor” to 0.1, “Use Marquart technique” check the Fudge factor 0.001; OK;
Yes+start;
Refinement converges with Rp~13.2%, R(obs)~ 11.2%.
“File → Structure → History”;
Select the parent structure cr2wo6;
“Tools → Special tools → Representation analysis for magnetic structures”
Repeat the same process for the second orthorhombic C centered Shubnikov space group
Cmm’m and use the default name i.e. cr2wo6_02 or any other;
Refinement converges with Rp~8.2%, R(obs)~ 5.4%.
Repeat the process also for Pnn’m’ and Pnn’m;
Use the same equation b=a as for the centered cell;
Pnn’m’ (allows ferromagnetic ordering) : Rp~13.2%, R(obs)~ 10.5%,7.7%, 54.6% for all,
nuclear and magnetic reflections, respectively.
Pnn’m (doesn’t allows ferromagnetic ordering) : Rp~8.2%, R(obs)~ 4.6%,4.7%, 3.3% for all,
nuclear and magnetic reflections, respectively.

9. Comparing of two best results Pnn’m and Cmm’m

[On the screen: Basic Jana window]
Up to now atomic coordinates were kept fixed.
Select the structure cr2wo6_02 (or any other) describing the Cmm’m model;
Right-click the icon of Refine and select the page “Various”
Select “Fixed commands” and disable the command “fixed xyz *”;
OK;OK; Yes+start;
Refinement converges with Rp~6.8%, R(obs)~ 2.1% for 19 profile and 12 structural
parameters
Repeat the process with Pnn’m
Refinement converges with Rp~6.7%, R(obs)~ 2.0%,2.0%, 2.3% for all, nuclear and
magnetic reflections, respectively – number of 19 profile and 10 structural parameters
Select the structure cr2wo6_02 (or any other) describing the Cmm’m model;
Double click on the icon “Plot structure”
Press the button “Atom filter” and select only Cr atoms;
Select “Use Vesta to draw the magnetic structure”

477 Jana2006 Cookbook, version October 2017

Select the structure cr2wo6_04 (or any other) describing the Pnn’m model;
Double click on the icon “Plot structure”
Press the button “Atom filter” and select only Cr atoms;
Select “Use Vesta to draw the magnetic structure”:

478 Jana2006 Cookbook, version October 2017

 The model in Pnn’m gives slightly better R values but the difference is so small that it is
difficult to make from it any final conclusion about the direction of the magnetic moment of
Cr atoms.

10. Orthorhombic strain for Pnn’m and Cmm’m

While the P centered cell are identical with that used for the parent tetragonal cell the C
centered cell is related to the original cell by equation:

𝒂𝒂𝐶𝐶−𝑜𝑜𝑜𝑜𝑜𝑜ℎ𝑜𝑜 = 𝒂𝒂𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡 + 𝒃𝒃𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡 𝒃𝒃𝐶𝐶−𝑜𝑜𝑜𝑜𝑜𝑜ℎ𝑜𝑜 = −𝒂𝒂𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡 + 𝒃𝒃𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡 𝒄𝒄𝐶𝐶−𝑜𝑜𝑜𝑜𝑜𝑜ℎ𝑜𝑜 = 𝒄𝒄𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡

The orthogonal ordering of magnetic moments can have some influence to the cell
parameters a and b which need not be same anymore. In order to test orthorhombic strain
without mixing it with the refined magnetic structure we can use le Bail profile fitting.

[On the screen: Basic Jana window]

Select the structure cr2wo6_02 (or any other) describing the Cmm’m model;
“File → Structure → Save as”;
Select name e.g. cr2wo6_02_cell; OK; Yes to the question “Do you want to continue with
the new structure?”;
Right-click the icon of Refine and select the page “Basic”;
Select “Make only profile matching (le Bail refinement)”;
Go to the page “Various”; Press the button “Equation” and deniable the equation b=a;
OK; OK; Yes;
Start “Edit profile”; Go to the page cell and activate there refinement of the b parameter
and modify slightly its value to avoid full overlaps;
OK; Yes to rewrite changed files;
Run refinement;
Refinement converges with Rp~6.5%,
Yes to open listing;
Press “Go to” and select “Changes overview”;

The cell parameters are: a=6.4849(23) and b=6.4850(23) which means that there is no strain
within the accuracy following from the refinement.

Repeat the procedure with Pnn’m model;

Refinement converges with Rp~6.4%,

479 Jana2006 Cookbook, version October 2017

 The cell parameters are: a=4.5876(3) and b=4.5837(3) which means that there is significant
difference connected with orthorhombic strain .

11. Final model Pnn’m

The refined magnetic moment of the Cr atoms is: (0.19(14),2.13(2),0) Bohr magnetons,
which means that the x component is negligible. In the final refinment we shall keep the x
component 0.

[On the screen: Basic Jana window]

Select the structure cr2wo6_04 describing the Pnn’m model;
“File → Structure → Save as”;
Select name e.g. cr2wo6_04_final; OK; Yes to the question “Do you want to continue with
the new structure?”;
Right-click the icon of Refine and select the page “Various”
Press the button “Fixed commands” and select “Set individual”;
Into the editing window “Atoms/parameters” write mx0[Cr1];
Add; OK;OK; Yes+start;
Refinement converges with Rp~6.7%, R(obs)~ 2.0%,2.0%, 2.2% for all, nuclear and
magnetic reflections, respectively – number of 20 profile and 10 structural parameters.

480 Jana2006 Cookbook, version October 2017

Example 13.1: TTB
Solution of a tetragonal-tungsten-bronze type structure from EDT (electron diffraction
tomography) data. Complete procedure from data reduction to structure refinement.

Revised: 04 March 2015

Chemical formula: K3Nb(5+x)O15

Electron diffraction data was measured on a transmission electron microscope Philips CM120
with a precession divice DigiStar. Accelerating voltage 120 kV.
Input data: directory dp with the set of experimental data
ttb.pts: input file for PETS – the data reduction of the beam stop
beamstop.xyz: file with the polygon defining the shape of the beam stop
ttb_start.cif_pets: result of the data processing. This file is provided so that this example can be
solved without the data processing part, and to warrant reproducibility of the second part.
File ttb.cif_pets is the output of the data reduction.

1. Preparing input file for PETS

Make sure you correctly installed PETS and ImageJ. Check Chapter “Installation notes” (page
7) for details. File ttb.pts is the input to PETS. The file was prepared by the software used for
the collection of the diffraction data, and most entries have been filled in. You have to enter
two parameters: the information about the beam stop and the size of reflections.
Open file ttb.pts in any plain-text editor (notepad, wordpad, notepad++, vim,…)
Open file dp/001.tif in ImageJ
In ImageJ go to “Image → Adjust → Brightness and Contrast”
Click the button “Auto” to adjust the constrast.
Go to “Image → Properties” and change the values of Pixel width and Pixel height from 0.005
to 1.
Measure the diameter of the reflections on the image.
There are several ways to measure the diameter of the reflections, but the simplest and
sufficient way is to place the cursor to the left of the reflection, remember or write down the
x-coordinate, then place the cursor to the right of the reflection, again read the x-coordinate
and subtract the two coordinates. The diameter should be large enough to encompass most
of the reflections, but not too large, to prevent overlaps. Do not take into account the broad,
weak TDS scattering around the strongest reflections! That would give you too large diameter.
In this case, value 25 is a good choice.
In ttb.pts, change the line “reflectionsize ??” to “reflectionsize 25”
In ImageJ, determine the coordinates of the tip of the beamstop (white cross on the
following image):

481 Jana2006 Cookbook, version October 2017

 You should get coordinates very close to 913 1014.
In ttb.pts change the line “beamstop ??(yes/no)” to “beamstop yes beamstop.xyz 913 1014”
File beamstop.xyz contains the coordinates of the polygon describing the beamstop relative
to a reference point. 913 and 1014 are the x- and y-coordinates of the reference point,
which is in this case the tip of the beamstop. The reference point can change from one
experiment to another, and must be therefore determined by the user.
Save and close ttb.pts

2. Run PETS and prepare files for indexation

Make sure ImageJ is not running. If it is, close ImageJ.
Run PETS with the file ttb.pts
There are several ways of running PETS. See installation notes for detailed description. A
window with initial menu of PETS opens, and ImageJ starts displaying a black image with a
white circle.
In the window of PETS, type p[Enter]
The peak hunting procedure starts. ImageJ sequentially displays diffraction images with
circles at points found by PETS as significant peaks. Note that PETS follows the position of
the primary beam, and corrects for small shifts. Note also the significant diffuse streaks.
Although they seem very intense, it is just due to the enhanced contrast of the images.
Compared to the Bragg intensities, the diffuse streaks are weak and do not disturb the
structure solution.
At the end, the evolution of the x- and y-coordinates of the estimated primary-beam position
is displayed (b&w inverted):

[On screen: main menu of PETS]

type r[Enter]
As the first step after peak hunting, PETS needs to refine the azimuthal angle between the
horizontal axis and the projection of the tilt axis. This angle depends on experimental
parameters, and needs to be refined for each data set. ImageJ displays a cylindrical projection
of the difference space of the extracted peak positions. For correct azimuthal angle, the image
contains sharp peaks aligned on sinusoidal curves. This step provides a first estimate of the
data quality. At the end of the refinement, you should get message:

482 Jana2006 Cookbook, version October 2017

 Final values of the parameters:
omega : 21.591
and see this image (b&w inverted):

[on screen: main menu of PETS]

type a[Enter]
PETS starts the analysis of the located peak positions. In the first step, distance distribution
between peaks in the image plane is analyzed. A sorted plot of inter-peak distances is
displayed (white curve on screen, black on the image below) with its derivative (grey curve):

For good quality data set, the white curve contains distinct steps, and, consequently, the grey
curve has distinct peaks.
Press Enter
In the next step an autoconvolution of the diffraction pattern (difference space) is analyzed,
and the groups of peaks in the autoconvolution (clusters) are replaced by the cluster centers.
Again, a distance distribution is displayed, and clear-cut steps on the curve indicate a well-
defined lattice.
Press Enter
Now the first step is finished and PETS prepared files needed for unit cell indexing. Keep the
window of PETS open, you will need it later!

3. Indexing with Jana2006

Start Jana2006
By default, Jana2006 starts with the last structure open. It may be cumbersome to change the
directory every time. You may create a dummy (empty) file ttb.m40 in the example directory,

483 Jana2006 Cookbook, version October 2017

and use the command “Open with” from the file-manager context menu to launch Jana2006
directly in this directory.
“Tools → Graphics → Indexing”
Locate and open file ttb.clust by double clicking it
A 3D view of the reciprocal lattice opens. You can rotate the cloud by mouse. Use buttons
and to switch between free rotation and rotation around z-axis. Rotate with the lattice
and try to determine the crystal system and the special axes of the lattice.
You should come to the conclusion that the lattice is tetragonal (warning – the lattice has a
few projections that look tetragonal, but are not exactly tetragonal). The following image
shows the view along c*:

IMPORTANT: before you continue, make sure you have the same orientation of the lattice on
your computer as displayed on the picture above (check the direction of the axes on the right).
If not, you may obtain slightly different results in the rest of the tutorial. While not necessarily
wrong, it makes following this tutorial more difficult.
Click the button in the lower-left corner
Click, hold the mouse button, and drag the marker to obtain the result on this image:

484 Jana2006 Cookbook, version October 2017

 Note that the vertical red line is on the tenth vertical row of peaks. By this action you defined
the direction and length of the (direct-space) basis vector a. Note: if you make a mistake and
need to redo this step, make sure that the radio button a is selected under the main window,
otherwise you will modify another cell parameter.
Repeat the action, this time drag the marker up vertically to define the vector b.
Both vectors should have lengths close to 12.8 Å, and angle gamma is close to 90°.
Click “View direction”, keep default selection “a”, Click OK
The reciprocal space is now viewed along a, and b points to the right.
In analogy to the procedure above, select the direction and length of the vector c (c points
up in this view).
When you are finished, at the bottom of the window you should see cell parameters close
(within a fraction of angstrom and degree) to these values:

“Refine cell → Refine UB+cell”

The orienting matrix is refined. Note that 90.3% of all peaks were indexed.
OK; Quit
A file ttb.smr was now produced by Jana2006. This file contains the orientation matrix and
the unit cell parameters.
Close Jana2006

4. Integrate the intensities with PETS

Return to the window of PETS.
Type u[Enter]
PETS reads the orienting matrix from the smr file.
Type i[Enter]
The integration procedure starts. ImageJ displays images with circles drawn at positions
predicted from the orienting matrix. A small circle inside large circle means that a peak with
I>3𝜎𝜎(I) was found. The small circle is located on the maximum of the intensity. A diamond
means that a reflection is expected on upcoming frame(s), square indicates reflection present
on previous frame(s).

485 Jana2006 Cookbook, version October 2017

 When the integration finishes, files ttb.cif_pets and ttb.hkl are created. Both can be used to
import the data into Jana and to solve the structure.

5. Create reciprocal-space sections

In PETS, Type h[Enter]
Type y[Enter]
You just asked PETS to scan all images and reconstruct sections through reciprocal space at
planes hk0, hk1, hk2, h0l, h1l, h2l, 0kl, 1kl, 2kl. We will use these sections as an aid to
determine the space group.
When the reconstruction finishes, type q[Enter] to close PETS. You may also close ImageJ.

6. Create new structure

Important! The data-processing procedure is almost never perfectly reproducible. Small
differences in the indexing and cell refinement procedure may result in small differences of
integrated intensities. If you want to be sure that you can exactly reproduce the following part
of the tutorial, do not use the file ttb.cif_pets that you just created, but use the file
ttb_start.cif_pets provided with the example files (Rename the file to ttb.cif_pets). Using your
own cif_pets file is also possible, but your results may slightly differ from the results described
in this tutorial.
Start Jana2006
“File → Structure → New” opens a file manager
Left pane: locate the directory with input files
Right pane: double-click ttb
Right pane detects possible Jana files and shows one jobname for each group of files

7. Import Wizard
Select “Known diffractometer formats”; NEXT
Select “Pets electron diffractometer”
The file name automatically changes to ttb.cif_pets, which is a file produced by PETS
containing all important information, including the list of intensities.
NEXT
Leave all settings unchanged (note the wavelength of 120kV electrons!); NEXT
Leave all settings unchanged; NEXT
The program reads 3777 reflections from the file
OK
For absorption correction select “None or done before importing”; NEXT
FINISH; OK; YES to accept the data set
You just read in the cell parameters and the intensity list. If you do not have the data in the
cif_pets format, you can read them in from a general hkl file using the option “reflection file
corrected for LP and absorption” at the beginning of the import wizard. You then have to input
the radiation type, wavelength and cell parameters by hand.

8. Symmetry Wizard
NEXT to close the information window and start the wizard
Symmetry Wizard can be started separately by “File → Reflection file → Make space group

486 Jana2006 Cookbook, version October 2017

 test. The default settings are prepared for x-ray diffraction. For electron diffraction we have
to increase the tolerances. Change the tolerances according to this screenshot:

Leave other settings default; NEXT; OK

[On the screen: Select Laue symmetry]
Select tetragonal Laue class 4/mmm; NEXT
The internal R-values would be rather high for x-rays, but are normal for electron diffraction
data, especially if measured in selected area mode.
[On the screen: Select cell centering]
Select primitive unit cell; NEXT
[On the screen: Select space group]
The window shows possible space groups.
The list of the strongest reflections contradicting the selected space group can be displayed
by “Details” button.
It is difficult to select the correct space group. In particular, presence or absence of the b-glide
is hard to asses, knowing that kinematically absent reflections may have significant intensities
due to dynamical diffraction effect. Open the file h0l.bmp and h1l.bmp, and compare the
intensities of reflections with h=2n (which should be absent in the h0l plane, but present in
h1l). You will see that despite a number of violations, the b-glide absences are nevertheless
clearly present in the h0l plane in comparison with the h1l plane. Space group with b-glide is
thus a reasonable guess. Note that a wrong choice of the space group is not detrimental for
the structure solution, because we will use Superflip, which will make additional check for the
most probable space group.
Select P4/mbm; NEXT
Accept the space group in the standard setting; FINISH
Symmetry is saved in file ttb.m50.

9. Creating refinement reflection file

In this step program creates file ttb.m90 containing the data set merged by symmetry and
with discarded forbidden reflections. M90 will be used for refinement.
Proceed with the wizard, leaving all settings default

10. Structure Solution Wizard

OK to close the information window and start the wizard
[On the screen: window of Structure solution] Structure solution wizard can be executed
separately through “Run → Solution”. Jana2006 does not contain solution procedures, it calls
external programs. By default, Jana uses Superflip (using charge flipping as the solution
method). Superflip is distributed with Jana2006.

487 Jana2006 Cookbook, version October 2017

 In “Formula” textbox type just the list of chemical elements: K Nb O
Superflip does not need information about chemical elements but they will be required for
peak assignment based on electron density map calculated by Superflip. If Jana2006 is used
for peak assignment, a list of expected chemical elements is sufficient. If you know the
expected formula, you may enter it and use EDMA for peak picking (at the bottom of the
window). If the formula is correct, EDMA often gives cleaner and more complete
interpretations of the solution.
Click the checkbox next to “Use a specific random seed”.
Leave other settings default; Press “Run solution”
The specific random seed is used just to guarantee reproducibility of the results. For standard
work random seed need not be defined.
[On the screen: window of Superflip appears with iteration running. After ten runs, open the
listing of Superflip log file in “Open the listing”]
Superflip performed ten solution attempts, and picked the best one. Its characteristics are
displayed near the bottom of the listing:
Properties of the saved density:
Run Rvalue Peaks Symm. Der.SG
8 24.46 2.71 3.24 P4/mbm
The space group proposed by Superflip is the same as our initial guess. The space-group
selection is confirmed. Small symmetry agreement factor (column Symm.) indicated good
quality of the solution. In cases of doubts or difficult problems, it is worth checking the
statistics, especially the symmetry analysis, for each of the ten runs, and seeking for the most
probable one.
Press “Close” to leave the listing
Press “Accept last solution”
The program creates file ttb.m40 with atoms of the structure model

11. Verification and completion of the structure solution

Go to “Contour → New plot → OK → Run 3D maps“
VESTA opens with a density shown as isosurfaces, and an atomic model overlaid. Rotate the
view to inspect the structure. Do you think it is complete?
You should come to the conclusion that the model is correct, all atoms were assigned correct
chemical type, only one oxygen (between two atoms Nb2 is missing.
Decrease the isosurface level to see weaker features in the density (“Properties →
Isosurfaces”, set isosurface level to 1)
You can see that there is a peak between the two Nb atoms, it just was not interpreted by
Jana2006 as an atom. Take-home message: it is always worth checking the density (i.e. the
raw result from Superflip), because it may contain features not considered in the interpreted
structure model.
Quit VESTA (do not save the file), quit Contour.
Start “Parameters → Atoms → New → Peaks from last Fourier calculation”
A list of maxima found in the density is displayed. In the right pane, the distances of the
maximum to the nearest atoms are displayed. First eight maxima correspond to atoms
interpreted by Jana2006. Max9 is a ghost peak, too close to Nb1. Max10 is at a correct
distance from two Nb2 atoms – this is the oxygen atom missing in the model.
Click “Include selected peak”, type O* in the field for name; OK; Finish; Yes to include the new
atom in the model

488 Jana2006 Cookbook, version October 2017

 You may now check the structure in Diamond to convince yourself that the structure is
complete and chemically reasonable.

12. Initial Refinement

[On the screen: basic window of Jana]
Right-click the icon of Refine.
[On the screen: refinement options]
Change the number of cycles to 100; OK
Choose “Yes+start”
Refinement converges with R value about 20.0%. At this moment you may check the
refinement listing with a lot of information about the refinement procedure and results.

13. Adding missing, partially occupied atom

[On the screen: basic window of Jana]
Right click the icon Fourier
On the first card, select F(obs)-F(calc) – difference Fourier
Modify the setting on the second card to agree with this screenshot:

OK; Yes+Start; No; No

Jana just calculated a difference Fourier map.
View the map in VESTA: “Contour → New plot → OK Run → 3D maps”
You will see that the difference map is dominated by one positive peak in the middle of an
empty cage of oxygens. This position tends to be partially occupied in some TTB structures. In
this case, the partial atom is Nb.
Add the missing atom. The procedure is the same as when you added the missing oxygen
atom after the solution. This time, the first maximum in the list is the correct maximum.
When defining the atom, do not forget to define the name as Nb* and type as Nb. Define
“Occupancy reduction” 0.2
OK; Finish; Yes
Now we will allow the occupancy of the niobium atom to be refined:
Double click “Edit atoms”; double click Nb3; on card “Edit” check the check box next to the
parameter “ai”; OK; OK; Yes
Double click Refine
The model refines to R-value about 19.4%. The improvement is very small, because the
occupancy of the Nb3 atom is very small, too, actually less than 10%.
This data set is exceptionally good. Normally, partial occupancies of atoms cannot be refined
reliably from electron diffraction data. Even here the refined occupancy must be taken only
as a rough indicator.
At this moment the structure is essentially finished. You may try to refine anisotropic
displacement parameters of the heavy atoms. The R-value decreases to ~18.3%. Again, with

489 Jana2006 Cookbook, version October 2017

this data set such refinement is possible, and the resulting parameters are reasonable. In
many other cases, anisotropic displacement parameters refine to unrealistic or even non-
positive definite values.

The resulting structure

490 Jana2006 Cookbook, version October 2017

Example 13.2: Kaňkite
Solution of kaňkite – a hydrated Fe-bearing arsenate.

Revised: 3 December 2013

Expected chemical formula: Fe+++AsO4•3.5H2O

Electron diffraction data measured on a transmission electron microscope Philips CM120 with
precession device DigiStar. Accelerating voltage 120 kV.
Input data: directory dp with the set of experimental data,
kankite.pts: input file for PETS – the data reduction program
beamstop.xyz: file with the polygon defining the shape of the beam stop
kankite_start.cif_pets: result of the data processing. This file is provided so that this
example can be solved without the data processing part, and to warrant
reproducibility of the second part. File kankite.cif_pets is the output of the data
reduction.

1. Data reduction using PETS

The data processing part follows exactly the procedure described in example 13.1. Please
follow that tutorial. Here only key results are provided for your control, and comments are
provided for stages that require special attention.
Parameters filled in the prepared file kankite.pts:
reflectionsize 25
beamstop yes beamstop.xyz 927 1025
Refined azimuthal angle omega of the tilt axis:
omega : 11.752
Unit cell is monoclinic, C-centered. The following screenshot is the view of the reciprocal lattice
along the monoclinic axis b (a* horizontal, c* vertical, pointing down):

The direction and length of the b-axis is difficult to determine. The lattice is obscured by quite
a lot of noise. This noise comes from the diffuse nature of the reflections along b* and from
the mosaicity of the crystal. If such noise is present, it is amplified by the data processing
preceding the creation of the file .clust. In such cases it is useful to check the correctness of

491 Jana2006 Cookbook, version October 2017

 the cell determination in the file .xyz. This file contains only the peak positions in reciprocal
space, not their autoconvolution. It is thus cleaner, but much less complete.
Click “New plot”
Locate and open the file kankite.xyz
You will see that the plot is less complete, but the noise is less abundant.
Final refined cell parameters (refined against the peaks in kankite.clust):

Quit the indexing plugin to create file kankite.smr with the orienting matrix, and continue
data processing in PETS by typing u[Enter] to read in the orienting matrix, and i[Enter] to
integrate the intensities.
You will note significant overlap of integration boxes (white circles) in the images displayed
by ImageJ. This is, however, not a real problem, because the lattice is C-centered, and every
second reflection is systematically absent.
PETS produced file kankite.cif_pets, which will be imported in Jana2006 in the next step.
Type h[Enter] and y[Enter] ]to produce standard sections through reciprocal space.
With this step the data processing in PETS is finished.
Type q[Enter] to close PETS. You may also close ImageJ.

2. Create new structure

Important! The data-processing procedure is almost never perfectly reproducible. Small
differences in the indexing and cell refinement procedure may result in small differences of
integrated intensities. If you want to be sure that you can exactly reproduce the following part
of the tutorial, do not use the file kankite.cif_pets that you just created, but use the file
kankite_start.cif_pets provided with the example files (Rename the file to kankite.cif_pets).
Using your own cif_pets file is possible, but your results may slightly differ from the results
described in this tutorial.
Start Jana2006
“File → Structure → New” opens a file manager
Left pane: locate the directory with input files
Right pane: double-click kankite
Right pane detects possible Jana files and shows one jobname for each group of files

3. Import Wizard
Select “Known diffractometer formats”; NEXT
Select “Pets electron diffractometer”
The file name automatically changes to kankite.cif_pets, which is a file produced by PETS
containing all important information, including the list of intensities.
NEXT
Leave all settings unchanged; NEXT
Leave all settings unchanged; NEXT
The program reads 5921 reflections from the file
OK
For absorption correction select “None or done before importing”; NEXT FINISH; OK; YES to
accept the data set
You just read in the experimental parameters, cell parameters and the intensity list.

4. Symmetry Wizard
NEXT to close the information window and start the wizard

492 Jana2006 Cookbook, version October 2017

 Symmetry Wizard can be started separately by “File → Reflection file → Make space group
test. The default settings are prepared from x-ray diffraction. For electron diffraction the
tolerances should be increased. Change the tolerances according to this screenshot:

Leave other settings default; NEXT

Jana detected a seven times larger unit cell with orthorhombic metric. This is not likely to be
the correct unit cell.
NEXT
[On the screen: Select Laue symmetry]
Select monoclinic Laue symmetry; NEXT
The internal R-value of 13.21 % is very good for electron diffraction data.
[On the screen: Select cell centering]
From the table the fact that the cell is C-centered is not obvious at all. This is due to the overlap
of the reflection positions during the integration. But we know the centering from the data
reduction and indexing step. You may convince yourself about the correctness of the centering
by inspecting the reciprocal space sections hk0, hk1 and hk2 generated by PETS.
Select C-centering; NEXT
[On the screen: Select space group] The window shows possible space groups. It is not obvious,
if the c-glide is present or not. Out of 60 reflections forbidden by the c-glide 33 have observed
intensity. Again, you should use the reciprocal-space sections to check for the c-glide.
However, do not use the h0l section! Due to the short b* axis and large mosaic spread, this
section contain “ghosts” from neighboring sections, and the systematic absences are
obscured. Inspection of the 0kl and 2kl section shows clearly that the reflections of the type
h0l: l=2n are absent, and c-glide is thus most likely present.
At this point we cannot decide, which of the choices Cc or C2/c is correct. Centrosymmetric
space groups are, in general, more frequent, and we will use it as a first guess.
Select C2/c; NEXT
Note that a wrong choice of the space group is not detrimental for the structure solution,
because we will use Superflip, which will make additional check for the most probable space
group.
Accept the space group in the standard setting; FINISH
Symmetry is saved in file kankite.m50.

5. Creating refinement reflection file

In this step program creates file ttb.m90 containing the data set merged by symmetry and
with discarded forbidden reflections. M90 will be used for refinement.
Proceed with the wizard, leaving all settings default
The internal R-value decreased from 13.21% To 10.02% due to the exclusion of the
systematically absent reflections.

6. Structure Solution Wizard

493 Jana2006 Cookbook, version October 2017

 OK to close the information window and start the wizard
[On the screen: window of Structure solution]
Structure solution wizard can be executed separately through “Run → Solution”. Jana2006
does not contain solution procedures, it calls external programs. By default, Jana uses
Superflip (using charge flipping as the solution method). Superflip is distributed with
Jana2006.
In “Formula” textbox type the expected formula: Fe2 As2 O15
In “Formula units” textbox type 4
This value gives a reasonable value of density: 3.01 g/cm3
Click the checkbox next to “Use a specific random seed” and leave the value 111.
The specific random seed is used just to guarantee reproducibility of the results. For standard
work, random seed need not be defined.
Click the radio button “EDMA – take as it is”
EDMA is an external program for interpretation of electron densities. It can take the chemical
formula and interpret the density using this formula.
Leave other settings default; OK
[On the screen: window of Superflip appears with iteration running. After ten runs, listing of
Superflip log file is displayed]
Superflip performed ten solution attempts. Check the space groups derived by Superflip from
each of the attempts. The most frequently proposed space group is Cc (eight of ten attempts).
Inversion center and two-fold axis have much worse symmetry agreement factors. It is likely
that the correct space group is Cc and not C2/c.
Press CLOSE to leave the listing
Press “Discard the result”
Back in the structure solution window, change the space group to Cc and recreate the
refinement reflection file
OK
[On the screen: Superflip runs again, this time creating output in Cc. EDMA is used to interpret
the density]
“Accept the result”
The result is a file kankite.m40 with atoms of the structure model.

7. Verification and completion of the structure solution

Go to “Contour → New plot → OK → Run 3D maps“
VESTA opens with a density shown as isosurfaces, and an atomic model overlaid. Rotate the
view to inspect the structure. Try to recognize the structural motifs. The main expected motifs
are AsO4 tetrahedra and FeO6 octahedra. The analysis will be simpler if you display the bonds
between atoms:
Go to Edit → Bonds, and set up the options according to this screenshot:

494 Jana2006 Cookbook, version October 2017

 It is also useful to show only positive isosurfaces (Properties > Isosurfaces). Viewing the
structure solution in VESTA has the advantage that you can directly check, if the raw density
was correctly interpreted. In this case, atom As1 should be Fe, and Fe1 should be As. You may
inspect the model also in Diamond. You should come to the conclusion that the model is
almost complete. Two oxygen atoms are missing (around Fe2) and five oxygen atoms in the
list (O10, O11, O12, O14 and O15) are ghost peaks.

8. Delete ghost atoms and change atomic types

Double click icon Edit atoms
Select atoms O10, O11, O12, O14 and O15 (Click while holding Ctrl)
“Action → Delete atoms → Yes to all”
Select atom As1
“Action → Edit/Define atoms → Type → Fe”
Repeat the assignment process to atom Fe1 (type As)
OK
“Select all → Action → Rename selected atoms to “atom_type”+number”
OK; Yes

9. Structure refinement
[On the screen: basic window of Jana]
Right-click the icon of Refine.
[On the screen: refinement options]
Set Number of cycles to 100; OK Choose “Yes+start”
Refinement converges to an R value about 19%

10. Adding the missing atom[On the screen: basic window of Jana]
Right click the icon Fourier
On the first card, select F(obs)-F(calc) – difference Fourier
Modify the setting on the second card to agree with this screenshot:

495 Jana2006 Cookbook, version October 2017

 OK; Yes+Start; No; No
Jana just calculated a difference Fourier map.
View the map in VESTA: “Contour → New plot → OK → Run 3D maps”
You will see that the difference map is dominated by two positive peaks around the atom Fe2.
These maxima correspond to the two missing atoms.
Close VESTA and return to the main window of Jana2006
“Parameters → Atoms → New → Peaks from last Fourier calculation”
A list of maxima found in the density is displayed. In the right pane, the distances of the
maximum to the nearest atoms are displayed. The first two maxima correspond to the missing
oxygen – they have the right distance to Fe2.
Select Max1, click “Include selected peak”, type O* in the field for name; OK;
Select Max2, click “Include selected peak”, type O* in the field for name; OK;
Finish; Yes to include the new atoms in the model
Double click the icon Refine to refine the completed model.
The structure refines to 15.89%. At this moment the refinement is finished. It is possible to
refine anisotropic ADP parameter of the atoms. The R-value drops to 11.63%, but the ADPs
are unrealistic – they have an extremely large component along b. This effect comes from the
data quality – reflections with high k-index are systematically weakened.

11. Identification of OH and H2O groups

The data quality is by far not good enough to see directly the hydrogen atoms. However, we
can estimate the positions of OH and H2O groups by the coordination and bond-valence sums
of the oxygen atoms.
Right-click icon Dist
Click button “Define coefficients for bond valences”
Define pair of atoms Fe, O; Click “From file → Fe3+ O2- (fourth in the list) → OK → Add”
Repeat for pair As – O, select the entry As5+ O2- from the list → OK → Add
OK;
[On screen: main dialog window of the Dist module]
Uncheck checkbox “dmax derived from atomic radii and typical distances”
OK; Yes+start; Yes to open the listing
At the end of the listing you will find the following table of bond-valence sums:

Bond valence for : Fe1 2.48(6)

Bond valence for : As1 4.32(12)
Bond valence for : As2 5.46(19)
Bond valence for : Fe2 2.33(6)
Bond valence for : O1 1.42(8)
Bond valence for : O2 1.86(8)
Bond valence for : O3 1.53(7)

496 Jana2006 Cookbook, version October 2017

Bond valence for : O4 1.46(4)
Bond valence for : O5 2.08(15)
Bond valence for : O6 0.99(5)
Bond valence for : O7 0.35(3)
Bond valence for : O8 0.90(5)
Bond valence for : O9 0.40(2)
Bond valence for : O10 1.53(5)
Bond valence for : O11 0.326(18)
Bond valence for : O12 1.75(8)
You can see three types of oxygen atoms. Most of oxygens have bond-valence sums above
1.4, two around 0.9 (O6 and O8) and three very low – below 0.5 (O7, 09, 011). O6 and O8
correspond to OH groups, and O7, O9, O11 are water molecules coordinating Fe atoms. This
is a view of the structure with OH and H2O groups labeled:

Thus, the published formula is not correct. The correct formula is

Fe2(OH)(AsO4)(AsO3OH).3H2O or Fe2As2O7(OH)2.3H2O

497 Jana2006 Cookbook, version October 2017

498 Jana2006 Cookbook, version October 2017
 Example 13.3: Orthopyroxene
Refinement of atomic occupancies from electron diffraction data using
dynamical diffraction theory
Revised: October 2015

Background information:
Orthopyroxene (opx) is a rockforming mineral – iron-magnesium silicate (Fe,Mg)2Si2O6,
Pbca, a=18.268, b= 8.868, c=5.202. It contains two symmetry-independent octahedral sites
with Mg and Fe. The distribution of Mg and Fe in these two sites depends on the
crystallization conditions, and it can therefore be used to infer information about the
formation of the rock.
Orthopyroxene often forms only microcrystalline aggregates or intergrowths with other
minerals. It is thus desirable to determine the occupancies of Fe and Mg on the two sites
from individual micro- or nano-crystals of opx.

Electron diffraction data were measured on FEI Tecnai G2 20 operated at 200 kV with
precession device DigiStar.

Input data: opx_001.tif: the experimental diffraction pattern, zone axis [001] of
orthopyroxene.
opx_001.pts: input file for PETS with experimental information
opx_xray.cif: cif file with the model of the opx structure refined against x-ray
data set.
opx_001_dyn_start.cif_pets: result of data processing with PETS. This file is
provided so that this example can be solved without the data processing part.

1. Data reduction using PETS

File opx_001.pts contains instructions for PETS. Open the file and check the instructions.
Compared to tomographic data, the input is much simpler in this case – the tilt-axis azimuth
need not be defined, and the orientation angles can be set to 0.
Note the line “background 32768”. The tif file was obtained by conversion of a dm3 file
(Digital Micrograph format). By default, the zero level of such file is at 215=32768. The
byword background 32768 corrects for this shifted zero level.

Start PETS with the file opx_001.pts

Type p[Enter]
PETS will extract peaks from the image and save them in file opx_001.cor. Check the image
shown by ImageJ. The central part is shown here:

499 Jana2006 Cookbook, version October 2017

The encircled black cross shows the predicted position of the primary beam. Obviously, the
automatic location of the center failed. We have to provide the center explicitly.
Type q[Enter] to quit PETS.
Open the file opx_001.tif in ImageJ
Go to “Image → Properties” and change the values of Pixel width and Pixel height from
0.005 to 1.
Read the coordinates of the direct beam.
The coordinates are close to 1990 1393
Close ImageJ
Open file opx_001.pts in a plain-text editor (e.g. Notepad) and modify the line “center
AUTO” to “center 1990 1393”
Save and close opx_001.pts.
Start PETS again with the updated file opx_001.pts
Type p[Enter]
This time the center is located properly.
We will not use the peak-processing functions of PETS, because the amount of data is not
sufficient. We will use directly the file opx_001.cor produced by PETS after the peak hunting.
Start Jana2006
Go to “Tools → Graphics → Indexing”
Locate the example directory and select file opx_001.cor; OK
You will see a plot of located peaks. The peaks form only a 2D layer. We have to use a 2D
indexing option of the indexing plugin to index the layer.
Click the button “2d section ON”
Jana2006 finds the best plane through the peaks, and fixes the view on this plane.
Click on the “indexing” button in the lower left corner.
Now if you drag the mouse with left button pressed, you will move an arrow on the screen.
By this arrow, you should locate a reflection, of which you know the indices. This is a little
tricky, because the plane [001] contains systematically absent reflections (due to the a-
glide), and, in addition, a pseudo n-glide. You should locate the vector a according to the
following screenshot. Note that reflection 12 0 0 was selected:

500 Jana2006 Cookbook, version October 2017

Similarly, vector b is located as follows (reflection 0 6 0 selected, hence order 6):

If you get lost, open the original diffraction pattern in ImageJ and compare the distribution
of reflections with the peak positions in Jana2006.
Once the two vectors are defined, the selected 2D cell can be refined:
“Refine cell → Refine UB+cell”
You should obtain a result very close to the following 2D cell:

501 Jana2006 Cookbook, version October 2017

 OK to quit the refinement dialog
This cell must be converted to a 3D cell to produce an orientation matrix:
Click “2d section OFF”
A dialog for the 2D to 3D transformation appears
Type the cell parameters: 18.268 8.868 5.202 90 90 90
In the line labeled “(h,k,l) for 2d axial direction #1” type 1 0 0
In the line labeled “(h,k,l) for 2d axial direction #2” type 0 1 0
Click “Calculate”
You should obtain the following result:

OK; Quit
Now you have an orientation matrix that you can read back to PETS, and extract intensities.
[On screen: window with PETS, main menu]
Type u[Enter]
PETS read in the orientation matrix calculated by Jana2006.
Type d[Enter]
D stands for “integration for dynamical refinement”. PETS extracts all intensities up to
d*=1.4, and produces file opx_001_dyn.cif_pets.

2. Import structure in jana2006

3. Important! The data-processing procedure is almost never perfectly reproducible. Small
differences in the indexing and cell refinement procedure may result in small differences of
integrated intensities. If you want to be sure that you can exactly reproduce the following
part of the tutorial, do not use the file opx_001_dyn.cif_pets that you just created, but use
the file opx_001_dyn_start.cif_pets provided with the example files. Before continuing,
make a copy of the file and name it opx_001_dyn.cif_pets. Using your own cif_pets file is
also possible, but your results may slightly differ from the results described in this tutorial.
4. Now we will import the file opx_001_dyn.cif_pets in Jana2006.
Start Jana2006
“File → Structure → New”
In the left pane, locate the example directory

502 Jana2006 Cookbook, version October 2017

 In the right pane, double click the label opx_001_dyn
Select “Single crystal: known diffractometer format”; NEXT
Select “Pets – electron diffractometer”; NEXT; NEXT; NEXT; OK; NEXT; FINISH; OK; Yes
You have just imported the data from cif_pets using default options. At this point Jana offers
symmetry determination wizard, but we will escape the standard procedure, which is not
useful for us.
Cancel; Cancel to cancel the symmetry wizard and the production of reflection file.
Now we have to import the structure model from the provided CIF file:
“File → Import model from → CIF”
Locate file opx_001.cif and double click it
OK to accept the new model
Cancel to avoid production of the refinement reflection file
If we accepted to produce the refinement reflection file at this stage, it would be prepared
for kinematical refinement. We will use dynamical refinement.
5. Setting up dynamical refinement
[On screen: Main windows of Jana2006]
“Parameters → Electron diffraction”
You should see this dialog:

This dialog contains experimental information relevant for the dynamical refinement, and
information about the zone axis (or several zone axes, if there are more zone axes
imported). Most options are now blank, because the reflection file for dynamical refinement
has not yet been prepared.
Check the box next to “Use dynamic approach”
OK to create refinement reflection file
Follow the wizard and accept default choices
[On screen: Electron diffraction dialog]
Now we have to find initial values for the sample thickness, scale factor and orientation. As
a first quick check, we will calculate only the thickness curve, using fixed orientation. This

503 Jana2006 Cookbook, version October 2017

calculation is much faster that the optimization of orientation, and it provides first
estimated of the data quality.
Click “except of scale optimize also: Thickness  Run optimization”
Jana2006 generates reflections and calls Dyngo.exe. Dyngo.exe calculates the reflection
intensities using dynamical diffraction theory.
[On screen: Electron diffraction dialog]
Next to the indication of the zone number you see the best R-value: R(all)=8.1174%. This is a
very promising value and indicates that already now the fit of the model to the data is quite
good.
Click “Show thickness plots”
You should obtain the following plot:

This plot shows a clearly defined minimum at the thickness 400 Å. This together with the low
R-value shows that the starting point for the refinement is reasonable.
Quit the display of the thickness plot
Click “except of scale optimize also: orientation  Run optimization”
The orientation of the experimental pattern is never perfect. Dyngo finds the parameters of
the misalignment of the zone axis from the perfect orientation predicted by the orientation
matrix. This calculation requires repeated evaluations of reflection intensities and may take
some time. You may follow the progress of the simplex algorithm in the window of Dyngo.
After the optimization, the zone axis parameters look like this:

The R-value decreased only slightly to 7.9963%. Value 0.15 of EDtheta means that the
diffraction pattern was recorded at an angle of 0.15 degrees away from the perfect
504 Jana2006 Cookbook, version October 2017
 orientation. EDPhi is the azimuth of the tilt axis. Given the large precession angle, this value
is within the precision of the method, and means that the pattern was essentially perfectly
oriented.
Quit the Electron diffraction dialog and confirm changes in the m42 file.

6. Dynamical least-squares refinement

We will first refine only the scale and thickness.
“Right-click Refine → Various → Fixed commands”
In the line labeled “Atoms/parameters” type * (star)
Click the arrow next to “Add”
You have fixed all structure parameters of all atoms. An exception is the atomic occupancy,
which is not fixed by this command, but we did not explicitly allow its refinement either, and
therefore the occupancies will remain, for the moment, fixed, too.
OK; OK; Yes+Start
Refinement converges smoothly to an R-value of 7.28%.
Now we will allow the refinement of occupancies.
“Right-click Refine → Various → Restrictions”
In the line labeled Atoms/molecules type “Fe1 Mg1”
Click radio button next to “Occupancies: keep overall sum”
Click the arrow next to “Add”
Repeat the action for the pair “Fe2 Mg2”
OK; OK; Yes to save the changes, but do not start the refinement
We have now restricted the occupancies of the two atoms to sum to 1. Only one occupancy
for each pair will be refined.
Now we have to allow the refinement of occupancies of Fe1 and Fe2:
Double click the icon “Edit atoms”
Double click atom Fe1
On the card “Edit” check the box next to the parameter “ai”; OK
Repeat for atom Fe2
OK to close the Edit atoms dialog
Double click Refine to start refinement
Refinement converges smoothly. R-value decreases only very little to 7.25%.
Refined occupancies are o(Fe1)=0.050(11), o(Fe2)=0.52(11). The occupancies refined from
the same sample with x-ray data are o(Fe1)=0.035(3), o(Fe2)=0.552(3). In general, the
refinement against electron diffraction data is in a good agreement with the refinement
against bulk material using x-rays.
If you wish, you may convince yourself that even from more distant starting values of the
occupancies the refinement will converge to the same values. You may, however, need to
use smaller damping factor (right-click Refine → Damping factor) to avoid oscillation or
locking in a false minimum, if you start from too distant values.

7. Kinematical least-squares refinement

We will perform a kinematical refinement for comparison.
[On screen: main window of Jana2006]
”File → Structure → Save as”
Type “opx_001_kin” in the window under the right pane; OK; Yes
We saved the structure under another name to preserve the dynamical refinement.

505 Jana2006 Cookbook, version October 2017

 “Parameters → Electron diffraction”
Uncheck the box next to “Use dynamic approach”; Yes to save changes
Accept the proposal to recreate the refinement file and follow the instructions. Leave all
settings default.
Double click “Refine” to start refinement
With a few clicks we switched from dynamical to kinematical refinement. The refinement R-
value is 31.03%, and the refined occupancies are o(Fe1)=0.280(87) and o(Fe2)=0.435(90).

506 Jana2006 Cookbook, version October 2017

Example 13.4: Silicon
Complete procedure from structure solution to dynamical structure refinement
from three zone-axis patterns.

Revised 11 December 2013

Input data: si_ZAPs.cif_pets: a cif_pets file containing the orientation matrix, positional angles
and intensity information from three zone-axis patterns collected on the same sample – a
wedge-shaped silicon single crystal. The file was obtained from raw diffraction patterns using
the procedure analogical to example 13.1.
1. Import data to Jana2006
Start Jana2006
“File → Structure → New”
In the left pane, locate the example directory
In the right pane, double click green label si_ZAPs
Select “Single crystal: known diffractometer format”; NEXT
Select “Pets – electron diffractometer”; NEXT; NEXT; NEXT; OK; NEXT; FINISH; OK; Yes
You have just imported the data from cif_pets using default options.
As a next step Jana starts the symmetry determination wizard. The data set we have is not
really complete, but we may nevertheless attempt to proceed with the wizard:
Next to start the wizard
On the next screen, set up the tolerances according to this screenshot:

Next; OK;
You now look at a table of Rint values for various Laue classes. The cubic Laue classes have
acceptable Rint values, and there is little doubt the structure is indeed cubic.
Select Laue class m-3m; Next
The data were already integrated assuming F-centering, therefore there are zero reflections
matching the F-centering absences in the data set.
Select centering F; Next
On the next screen a list of possible space groups is presented. For the space group Fd-3m, 7
out of 24 forbidden reflections have I>3σ(I). However, looking at the details (click the button
“Details”) you will find that these reflections are in fact only two - 0 0 2 and 0 0 6. The d-
glide is thus likely to be present. Indeed, the space group of silicon is Fd-3m.
Select “Fd-3m”; Next; Finish
In the next step we prepare a “standard” reflection file suitable for structure solution.
Next; OK; OK
[On screen: the window with three radio buttons for the choice of the handling of scales]
Jana treats can data from each zone axis on a separate scale. For the structure solution and
kinematical refinement, we will, however, assume that the data are on the same scale. The
number of common reflections is insufficient to determine a reliable common scale.

507 Jana2006 Cookbook, version October 2017

 Select “Ignore scale flags”; Next; Next; OK; Finish
2. Structure solution and kinematical refinement
[On screen: Information window explaining that the next step is structure solution]
OK to open the structure solution dialog
Type Si in the dialog box for formula; change the number of formula units to 8
Leave all settings default
OK to start solution by Superflip
Superflip runs and produces a structure solution.
When Superflip exits, accept the solution
Plot structure in Diamond to make sure it has been correctly solved
Right-click Refine; change number of cycles to 100; OK; Yes+start
Jana2006 refines the structure using kinematical calculation of diffracted intensities. The
structure refines to the following values (note that there are only 13 symmetry independent
reflections!):

Open the refinement listing (Edit → View of Refine) and check the statistics.
The displacement parameter of silicon refines to -0.0106.
3. Origin shift
The default origin of the space group Fd-3m is at the point -43m, which does not contain
inversion center. For the dynamical refinement it is desirable to have the origin in the
inversion center, if possible. We will now shift the origin of the unit cell to the inversion
center:
[On screen: Main window of Jana2006]
“Tools → Transformations → Origin shift”
Type “1/8” in all three fields; OK
4. Initiation of dynamical refinement
[On screen: Main window of Jana2006]
The displacement parameter of silicon did not refine to physically meaningful value. Before
initiating the dynamical refinement it is good to reset it to a more correct value. It is not
strictly necessary in this example, but it is generally a good practice.
Double click the icon Edit atoms
Double click Si1
On the card Edit, change Uiso to 0.005
OK; OK; Yes
[On screen: Main window of Jana2006]
“Parameters → Electron diffraction”

508 Jana2006 Cookbook, version October 2017

You will see the following dialog:

Default excitation error is too small for the current data – it must be increased:
Change Maximal excitation error to 0.03
Check the box next to “Use dynamic approach”
OK to create refinement reflection file
Follow the wizard and accept default choices
[On screen: Electron diffraction dialog]
Now we have to find initial values for the sample thickness, scale factors and orientations of
all three zones. As a first quick check, if the data quality is acceptable, we will calculate only
the thickness curve, using fixed orientation. This calculation is much faster that the
optimization of orientation, and it provides a first estimate of the data quality.
Click “Optimize thickness and scale”
Jana2006 generates reflections and calls Dyngo.exe for each zone separately. Dyngo.exe
calculates the reflection intensities using dynamical diffraction theory. Wait until Dyngo.exe
terminates.
[On screen: Electron diffraction dialog]
Click “Show thickness plots”
Go through the thickness plots. You should see three curves with more-or-less well defined
minima between 500 and 1000 Å. The best R-values are around 15%. This is a good start,
but it is still not satisfactory. It is necessary to optimize the orientations:
Quit the display of the thickness plot
Click “optimize thickness, scale and orientation”
This option optimizes all three parameters simultaneously, using a simplex algorithm. This
calculation requires repeated evaluations of reflection intensities and is thus much longer
that just the thickness optimization. Wait unil all three runs of Dyngo finish.
[On screen: Electron diffraction dialog]
Check the refined tilt angles. EDtheta is the azimuth of the tilt, EDphi is the amplitude. You
will see that the refined misorientations range between 0.13 and 0.39 degrees – very
reasonable values.
Click “Optimize thickness and scale”

509 Jana2006 Cookbook, version October 2017

 Repeat the thickness orientation to view the updated thickness curves. When viewing the
thickness curves again, you will see much better defined minima.
However, for zone 3 you will find the following curve:

There is a clear and reasonable minimum at t=350 Å, but the lowest R-value is found for the
thickness of 2000 Å. It is necessary to manually reset the thickness to 350 Å before the
refinement.
Quit; Locate zone 3 and set the thickness according to the following screenshot:

The data are now checked and ready for least-squares refinement.
OK; Yes to exit the Electron diffraction dialog
5. Structure refinement
[On screen: Main window of Jana2006]
The dynamical refinement tends to be unstable with default damping factor, it is thus better
to decrease it:
Right click “Refine”
Set damping factor to 0.2
OK; Yes+Start
The dynamical refinement starts. You may follow the progress in the window of Jana2006.
You may also interrupt the refinement or change the parameters the same way as for the
standard refinement.

The refinement converges after 21 cycles to R(obs)=6.8%. (Compare with 23% obtained with
kinematical refinement).

510 Jana2006 Cookbook, version October 2017

Uiso(Si1) refines to -0.0010(19). This is a much better value than the result of kinematical
refinement, but it is, of course, not satisfactory. A detailed investigation shows that even if
only one zone-axis at a time is used for the refinement, Uiso(Si1) remains slightly negative. It
is thus probably a problem related to the experiment. Unfortunately, the sample used for
the data collection was damaged, and we could not remeasure the data. Previous
measurements on the same samples yielded correct Uiso around 0.005, but these
measurements were single-zone only, and they are thus not so well suited for demonstration
of the complete procedure.

By repeating the optimization of orientation and rerefining the structure, the R-value can be
decreased even more to 5.0%. However, Uiso(Si1) remains negative.

511 Jana2006 Cookbook, version October 2017

512 Jana2006 Cookbook, version October 2017
Example 13.5: Ni2Si
Solution of nickel silicide Ni2Si by PEDT (precession electron diffraction
tomography) data. Complete procedure from data reduction to structure
refinement using the dynamical approach.
Revised: October 2015

Chemical formula: Ni2Si

Electron diffraction data was measured on a transmission electron microscope Philips CM120
with a precession divice DigiStar. Accelerating voltage of 120 kV.
Input data:
directory dp-150 with the set of experimental data (reflection images)
directory img with the set of experimental images from the crystal tilt
Ni2Si.pts: input file for PETS
Ni2Si_start.cif_pets: result of the data processing. This file is provided so that this
example can be solved without the data processing part, and to warrant
reproducibility of the solution and refinement part. File Ni2Si.cif_pets will be the
output of the data reduction.
Ni2Si_start_dyn.cif_pets: result of the data processing for dynamical refinement. This
file is provided so that the dynamical approach can be used without the processing
part, and to warrant the reproducibility of the refinement.

1. Preparing input file for PETS

Make sure you correctly installed PETS and ImageJ. Check Chapter “Installation notes”
(page 7) for details. File Ni2Si.pts is the input to PETS. The file was prepared by the
software used for the collection of the diffraction data, and most entries have been filled
in. You have to enter two parameters: the information about beam stop and the size of
reflections.
Open file Ni2Si.pts in any plain-text editor (notepad, wordpad, notepad++, vim,…)
Open file dp-150/001.tif in ImageJ
In ImageJ go to “Image → Adjust → Brightness and Contrast”
Click the button “Auto” to adjust the contrast
Go to “Image → Properties” and change the values of Pixel width and Pixel height from
0.005 to 1
Measure the diameter of the reflections on the image
There are several ways to measure the diameter of the reflections, but the simplest and
sufficient way is to place the cursor to the left of the reflection, remember or write down
the x-coordinate, then place the cursor to the right of the reflection, again read the x-
coordinate and subtract the two coordinates. The diameter should be large enough to
encompass most of the reflections, but not too large, to prevent overlaps. Do not take into
account the broad, weak TDS scattering around the strongest reflections! That would give
you a too large diameter. In this case, value 25 is a good choice.
In Ni2Si.pts, change the line “reflectionsize ??” to “reflectionsize 25”
This measurement was performed without beamstop (microdiffraction). In Ni2Si.pts
change the line “beamstop ??(yes/no)” to “beamstop no”

513 Jana2006 Cookbook, version October 2017

 Save and close Ni2Si.pts

2. Run PETS and prepare files for indexing

Make sure ImageJ is not running. If it is, close ImageJ.
Run PETS with the file Ni2Si.pts
There are several ways of running PETS. See installation notes for detailed description. A
window with initial menu of PETS opens, and ImageJ starts displaying a black image with
a white circle.
In the window of PETS, type p[Enter]
The peak hunting procedure starts. ImageJ sequentially displays diffraction images with
circles at points found by PETS as significant peaks. Note that PETS follows the position of
the primary beam, and corrects for small shifts. Note also the significant diffuse streaks.
Although they seem very intense, it is just due to the enhanced contrast of the images.
Compared to the Bragg intensities, the diffuse streaks are weak and do not disturb the
structure solution.
At the end, the evolution of the x- and y-coordinates of the estimated primary-beam
position is displayed (b&w inverted):

[On screen: main menu of PETS]

type r[Enter]
The next step after peak hunting is the refinement of the azimuthal angle between the
horizontal axis (x axis of the micrograph) and the projection of the tilt axis. This angle
depends on experimental setup, and needs to be refined for each data set. ImageJ displays
a cylindrical projection of the difference space of the extracted peak positions. A correct
azimuthal angle refinement results in the image containing sharp peaks aligned on
sinusoidal curves. This step provides a first estimation of data quality. At the end of the
refinement the omega angle should be close to 13°.
On the screen:
Final values of the parameters:
omega : 13.165

and see this image (b&w inverted):

514 Jana2006 Cookbook, version October 2017

 [on screen: main menu of PETS]

type a[Enter]
PETS starts the analysis of the peak positions obtained in the peak hunting procedure. In
the first step, distance distribution between peaks in the image plane is analyzed. A sorted
plot of inter-peak distances is displayed (white curve on screen, black on the image below)
with its derivative (grey curve):

For good quality data set, the white curve contains distinct steps, and, consequently, the
grey curve has distinct peaks.
Press Enter
In the next step an autoconvolution of the diffraction pattern (difference space) is
analyzed, and the groups of peaks in the autoconvolution (clusters) are replaced by the
cluster centers. Again, a distance distribution is displayed, and clear-cut steps on the curve
indicate a well-defined lattice.
Press Enter
Now the first step is finished and PETS prepared the files needed for unit cell indexing.
Keep the window of PETS open, you will need it later!

3. Indexing with Jana2006

Start Jana2006
By default, Jana2006 starts with the last structure open. It may be cumbersome to change
the directory every time. You may create a dummy (empty) file Ni2Si.m40 in the example
directory, and use the command “Open with” from the file-manager context menu to
launch Jana2006 directly in this directory.
“Tools → Graphics → Indexing”
Locate and open file Ni2Si.clust by double clicking it

515 Jana2006 Cookbook, version October 2017

A 3D view of the reciprocal lattice opens. You can rotate the cloud by mouse. Use buttons
and to switch between free rotation and rotation around z-axis. Rotate with the
lattice and try to determine the crystal system and the special axes of the lattice.
You should come to the conclusion that the lattice is orthorhombic. The following image
shows the view along c*:

IMPORTANT: before you continue, make sure you have the same orientation of the lattice
on your computer as displayed on the picture above (check the direction of the axes on the
right). If not, you may obtain slightly different results in the rest of the tutorial. While not
necessarily wrong, it makes following this tutorial more difficult.
Click the button in the lower-left corner and choose the radio button for a vector
setting (right to the button ).
Click, hold the mouse button, and drag the marker to obtain the result on this image:

516 Jana2006 Cookbook, version October 2017

 Set the order of the interplanar distance to 3.
Note that the vertical red line is on the third vertical row of peaks. By this action you
defined the direction and length of the (direct-space) basis vector a. Note: if you make a
mistake and need to redo this step, make sure that the radio button a is selected bellow
the main window, otherwise you will modify another cell parameter.
Repeat the action, this time drag the marker up vertically to define the vector b.
Vector a should have length close to 5.00 Å, vector b should have length close 3.73 Å,
angles alpha and beta should be close to 90°.

Click “View direction”, select “a”, Click OK

The reciprocal space is now viewed along a, and c points up.
In analogy to the procedure above, select the direction and length of the vector c.
When you are finished, at the bottom of the window you should see cell parameters close
(within a fraction of angstrom and degree) to these values:

“Refine cell → Refine UB+cell”

The orientation matrix is refined. Note that 44.9% of all peaks were indexed.
OK; Quit
The file Ni2Si.smr was now produced by Jana2006. This file contains the orientation matrix
and the unit cell parameters.
Close Jana2006

4. Integrate the intensities with PETS

Return to the window of PETS.
Type u[Enter]
PETS reads the orienting matrix from the Ni2Si.smr file.
Type i[Enter]
517 Jana2006 Cookbook, version October 2017
 The integration procedure starts. ImageJ displays images with circles drawn at positions
predicted from the orienting matrix. A small circle inside large circle means that a peak
with I>3𝜎𝜎(I) was found. The small circle is located on the maximum of the intensity. A
diamond means that a reflection is expected on upcoming frame(s), square indicates
reflection present on previous frame(s).
When the integration finishes, files Ni2Si.cif_pets and Ni2Si.hkl are created. Both can be
used to import the data into Jana and to solve the structure.

5. Create reciprocal-space sections

In PETS, Type h[Enter]
Type y[Enter]
You just asked PETS to scan all images and reconstruct sections through reciprocal space
at planes hk0, hk1, hk2, h0l, h1l, h2l, 0kl, 1kl, 2kl. We will use these sections as an aid to
determine the space group.
When the reconstruction finishes, type q[Enter] to close PETS. You may also close ImageJ.

6. Create new structure

Important! The data-processing procedure is almost never perfectly reproducible. Small
differences in the indexing and cell refinement procedure may result in small differences
of integrated intensities. If you want to be sure that you can exactly reproduce the
following part of the tutorial, do not use the file Ni2Si.cif_pets that you just created, but
use the file Ni2Si_start.cif_pets provided with the example files (Rename the file to
Ni2Si.cif_pets). Using your own cif_pets file is also possible, but your results may slightly
differ from the results described in this tutorial.
Start Jana2006
“File → Structure → New” opens a file manager
Left pane: locate the directory with input files
Right pane: double-click Ni2Si
Right panel detects possible Jana files and shows one jobname for each group of files

7. Import Wizard
Select “Known diffractometer formats”; NEXT
Select “Pets electron diffractometer”
The file name automatically changes to Ni2Si.cif_pets, which is a file produced by PETS
containing all important information, including the list of intensities.
NEXT
Leave all settings unchanged (note the wavelength of 120kV electrons!); NEXT
Leave all settings unchanged; NEXT
The program reads 614 reflections from the file
OK
For absorption correction select “None or done before importing”; NEXT
FINISH; OK; YES to accept the data set
You just read in the cell parameters and the intensity list. If you do not have the data in
the cif_pets format, you can read them in from a general hkl file using the option
“reflection file corrected for LP and absorption” at the beginning of the import wizard. You
then have to input the radiation type, wavelength and cell parameters by hand.

8. Symmetry Wizard
518 Jana2006 Cookbook, version October 2017
 NEXT to close the information window and start the symmetry wizard.
Symmetry Wizard can be started separately by “File → Reflection file → Make space group
Test”. The default settings are prepared for x-ray diffraction. For electron diffraction we
have to increase the tolerances. Change the tolerances according to this screenshot:

Leave other settings default; NEXT; OK

[On the screen: Select Laue symmetry]
Select orthorhombic Laue class mmm; NEXT
The internal R-values would be rather high for x-rays, but are low for electron diffraction
data.
[On the screen: Select cell centering]
Select primitive unit cell; NEXT
[On the screen: Select space group]
The window shows possible space groups.
The list of the strongest reflections contradicting the selected space group can be
displayed by “Details” button.
It is difficult to select the correct space group. In particular, presence or absence of the
glide is hard to asses, knowing that kinematically absent reflections may have significant
intensities due to dynamical diffraction effect. In this case the missing wedge leaves
almost no information in hk0 reciprocal-space section. Open images 0kl and 1kl and check
the intensities with k+l=2n+1 (which should be absent in the 0kl plane, but present in 1kl).
You will see that despite some violations and the missing information in the center of the
image, the n-glide absences are nevertheless clearly present in the 0kl plane in comparison
with the 1kl plane. Space group with n-glide is thus a reasonable guess.
Note that a wrong choice of the space group is not detrimental for the structure solution,
because we will use Superflip, which will make additional check for the most probable
space group.

Select Pnma; NEXT

Accept the space group in the standard setting; FINISH
Symmetry is saved in file Ni2Si.m50.

9. Creating refinement reflection file

In this step program creates file Ni2Si.m90 containing the data set merged by symmetry
and with discarded forbidden reflections. .m90 will be used for refinement.
Proceed with the wizard, leaving all settings default

10. Structure Solution Wizard

OK to close the information window and start the solution wizard
[On the screen: window of Structure solution] Structure solution wizard can be executed
separately through “Run → Solution”. Jana2006 does not contain solution procedures, it

519 Jana2006 Cookbook, version October 2017

 calls external programs. By default, Jana uses Superflip (using charge flipping as the
solution method). Superflip is distributed with Jana2006.

In “Formula” textbox type just the list of chemical elements: Ni Si

Superflip does not need information about chemical elements but they will be required for
peak assignment based on electron density map calculated by Superflip. If Jana2006 is
used for peak assignment, a list of expected chemical elements is sufficient. If you know
the expected formula, you may enter it and use EDMA for peak picking (at the bottom of
the window). If the formula is correct, EDMA often gives cleaner and more complete
interpretations of the solution.
Click the checkbox next to “Use a specific random seed” and enter the value 111.
Leave other settings default; “Run solution”.
The specific random seed is used just to guarantee reproducibility of the results. For
standard work random seed need not be defined.
[On the screen: window of Superflip appears with iteration running. After ten runs, listing
of Superflip log file is displayed]
Superflip performed ten solution attempts, and picked the best one. Its characteristics are
displayed near the bottom of the listing:
Properties of the saved density:
Run Rvalue Peaks Symm. Der.SG
10 51.46 2.23 1.37 Pnma
The space group proposed by Superflip is the same as our initial guess. The space-group
selection is confirmed. Small symmetry agreement factor (column Symm.) indicated good
quality of the solution. In cases of doubts or difficult problems, it is worth checking the
statistics, especially the symmetry analysis, for each of the ten runs, and seeking for the
most probable one.
Press “Close” to leave the listing
Press “Accept last solution”
The program creates file Ni2Si.m40 with atoms of the structure model

11. Verification of the structure solution

Go to “Contour → New plot → OK → Run 3D maps“
VESTA opens with a density shown as isosurfaces, and an atomic model overlaid. Rotate
the view to inspect the structure. Do you think it is complete?
You should come to the conclusion that the model is correct, correct elements were
assigned to all maxima.
Quit VESTA (do not save the file), quit Contour.

520 Jana2006 Cookbook, version October 2017

 You may now check the structure in Diamond to convince yourself that the structure is
complete and chemically reasonable.

12. Initial Refinement

[On the screen: basic window of Jana]
Right-click the icon of Refine.
[On the screen: refinement options]
Change the number of cycles to 100; OK
Choose “Yes+start”
Refinement converges with R value about 11.07%. At this moment you may check the
refinement listing with a lot of information about the refinement procedure and results.
As you can see, the refinement is already very good for electron diffraction. R values are
low, all the ADPs were positive definite. Refinement of anisotropic atomic displacements
would decrease the residue value (R=8.79%), but the ratio reflections per parameter is
already low.

The resulting structure

13. Preparation of reflection file for the dynamical refinement

Open Ni2Si.pts again. Type “o” to read the peak table, “u” to read the orientation matrix
and “d” to integrate the intensities for the dynamical refinement.
The integration for the dynamical refinement considers each diffraction image separately.
The result is a list with the hkl indices, the measured intensity with the respective sigma
and the number of the frame in which the intensity is present. PETS creates the file
Ni2Si_dyn.cif_pets with all these information, which will be used in Jana2006. Note that
the same hkl reflection might be present in several frames, with different intensity due to
the multiple scattering relative to different crystal orientation and thickness at this
condition.

521 Jana2006 Cookbook, version October 2017

 Quit PETS and go back to Jana2006. “StructureSave as” and save a new structure
Ni2Si_dyn for the dynamical refinement. Continue with the new structure.
This will be the structure with the dynamical refinement

“FileReflection fileImport/modify the reflection fileReimport”

Again, if you want to be sure that you can exactly reproduce the following part of the
tutorial, use the file Ni2Si_start_dyn.cif_pets provided with the example files. Rename this
file to Ni2Si_dyn.cif_pets.
Select the Ni2Si_dyn.cif_pets as the new reflection file. NEXT, NEXT, NEXT.
Now, Jana reads 4825 reflections from all the frames.
OK, NEXT, Finish, OK, Yes, Cancel.
It is not necessary to average the intensities, because the intensities will be used
independently for each frame during the dynamical refinement.

14. Starting the dynamical refinement

[On the screen: basic window of Jana]
“Parameters”, “Electron diffraction”.
[On the screen: Parameters for electron diffraction data]
Enable “Use dynamic approach”. OK, NEXT, OK, OK, OK, Finish.
Notice that now, 1673 reflections are symmetrically independent from the 4825
reflections read. From these, 12 were rejected because they violate the systematic
extinctions. You can see these information in the file .rre. In this same file you can see the
completeness of the data (63% for theta full (0.76 °).
[On the screen: Parameters for electron diffraction data]
The default values for the selection of the intensities are: g(max)=2, maximal excitation
error Sg(matrix)=0.01, maximal excitation error Sg(refine)=0.1 and Rsg=0.4.

15. Optimization of thickness

Enable “Except of scale optimize also: Thickness”, “Run optimizations”.
[On the screen: Intensity calculation for optimizing of thickness].
Dyngo calculates wR-factor for the thickness varying from 0 to 2000 Å.
“Show thickness plots”. Check the variation of the thickness of the crystal for the frames.
Quit.
Most frames have well defined wR-fator minima for one thickness. However, some frames
either have poorly defined minimum or the minimum is defined for very different
thickness. For example frames 13, 18, 22 and 44 have a plateau for thickness between 0
and 400 Å. Frame 36 have minimum at very high thickness (around 1500 Å). Check the
problematic frames. You will notice that for most frames there is no problem with the
reflection image, in this case, and for some of them 1 or 2 reflections are stronger the all
the rest. Notice also that the maximum R-factor is 17.64% for the last frame. You can
check the wR(all) going through the zones in the main window of “parameters for electron
diffraction data” or while seeing the graphs. The different thickness is not a problem now,
but if the frame would have some problem (the crystal mostly moved from the beam, for
example) it could be removed from the refinement.

16. Defining all the zones with the same thickness

“Define zones of equal thickness”. “Select all”, OK.
522 Jana2006 Cookbook, version October 2017
 The crystal is a nanowire and the thickness is mostly constant along the sample. However,
the thickness changes according to the tilt of the sample, because the nanowire lies
roughly perpendicular to the tilting axis of the sample holder. Then, it is possible to define
the same thickness for all the frames but still refine the variations of the thickness
according to the tilting of the sample.
There are two ways of defining the same thickness of the crystal for all the frames. You
can either estimate a reasonable thickness, based on the plots of wR versus thickness, and
type it for all frames, or run 0 cycle of refinement and average the thickness. It will be used
the first option here. From the plots the minima are well defined mostly for EDThick under
200 Å. Also based on the image of the crystal (it can be measured using imageJ) 150 (15
nm) for EDThick is a reasonable value.

“Select zones for editing Select all  OK” and type 150 for “EDThick”. OK, Yes.

Right-click on icon “Refine”, change the number of cycles to 100 and the damping factor
to 0.3 in order to avoid instability of the refinement.
The refinement runs 23 cycles, each cycle takes less than 2 minutes (depending on your
computer). You can see the details of the refinement in the file Ni2Si.ref (Edit/View View
of Refine). The refinement converges with R=7.79%. However, the ratio reflections per
parameter refined is around 5 (454/84) and it is advisable to increase it. It can be done by
changing RSg (selecting more reflections for the refinement).
Another important thing to check is the number of observed reflections per zone. Also in
the Ni2Si.ref file, you can see “R-factors for individual zones”, where it is shown the list of
zones with the respective R-factors and number of reflections.

17. Optimization of orientation

Save a new structure “Filesave asNi2Si_dyn_ori”, OK, Yes.
“Parameterelectron diffraction”, change the “Limit on RSg” to 0.75
Enable “Run optimizations except of scale optimize also: Orientation”, “Run
optimization”.
[On the screen: Intensity calculation for optimizing of orientation].
Check the EDTheta for the frames. As said before, the orientation of the crystal during the
measurement plays an important role in the dynamical scatterings. Typically, a deviation
of 0.5° is already considered high. The frames with high misorientation are: 8, 24, 32, 41,
60, 61, 62 and 73. If you check these frames, most have few reflections (because the unit
cell is small), the reflections are mostly weak (the nanowire is very thin) and some frames
have 1 or 2 peaks visibly stronger than the rest. These stronger reflections dominate the
optimization and hamper the correct estimative of orientation. Frames with EDTheta
larger than 0.5° will be removed from the refinement.

“Select zones for refinement”

Disable the frames 8, 24, 32, 41, 60, 61, 62 and 73 (Ctrl + left mouse button).

OK, OK, Yes. Double click on “Refine”.

The refinement converges with R=7.11% after 13 cycles. Since there are more reflections,
it is possible to refine the anisotropic displacement parameters.

18. Refinement of anisotropic atomic displacement parameters

523 Jana2006 Cookbook, version October 2017
Save the structure as Ni2Si_dyn_ori_aniso, “Edit atomsselect alldouble click in one
of the atomsADP parameter(s): harmonic”, OK, OK, Yes
[On the screen: Mains window of Jana2006].
Double-click on icon “Refine”.

Now the refinement converges with R= 6.52% after 15 cycles. If you check the ellipsoid
model, you will see that Si is elongated along a*, which was the noisiest direction when
defining the unit cell.

Final structure model:

524 Jana2006 Cookbook, version October 2017

Example 13.6:
Orthopyroxene from EDT
A complete procedure from structure solution to dynamical structure
refinement from electron diffraction tomography data.

Revised: 13 October 2015

Warning: this is a complex and time-consuming example. The computations in the last stages
may take up to several hours.

Input data: directory dp with the set of experimental data,

opx.pts: input file for PETS – the data reduction program
opx_start.cif_pets and opx_dyn_start.cif_pets: results of the data processing. These files are
provided so that this example can be solved without the data processing part, and to warrant
reproducibility of the second part. Files opx.cif_pets and opx_dyn_start.cif_pets is the output of
the data reduction.

1. Data reduction using PETS

The data processing part follows exactly the procedure described in example 13.1. Please
follow that tutorial. Here only key results are provided for your control, and comments are
provided for stages that require special attention. Be aware that the peak search and
integration take a long time, because the original images are rather large.
Parameters filled in the prepared file opx.pts:
reflectionsize 30
beamstop no

Refined azimuthal angle omega of the tilt axis:

omega : 57.922

Unit cell is orthorhombic, primitive. The following screenshot is the view of the reciprocal
lattice along the axis c (a* horizontal, b* vertical):

525 Jana2006 Cookbook, version October 2017

 Final refined cell parameters (refined against the peaks in opx.clust):

Quit the indexing plugin to create file opx.smr with the orienting matrix, and continue data
processing in PETS by typing u[Enter] to read in the orienting matrix, and i[Enter] to
integrate the intensities.
PETS produced file opx.cif_pets, which contains intensities suitable for structure solution and
kinematical refinement. We have to produce another intensity set for dynamical refinement:
In PETS type d[Enter]
PETS will make another integration, this time producing file opx_dyn.cif_pets.
Type h[Enter] and y[Enter] ]to produce a standard set of sections through reciprocal space.
The sections will be useful for confirming the space group.
With this step the data processing in PETS is finished.
Type q[Enter] to close PETS. You may also close ImageJ.

2. Create new structure

Important! The data-processing procedure is almost never perfectly reproducible. Small
differences in the indexing and cell refinement procedure may result in small differences of
integrated intensities. If you want to be sure that you can exactly reproduce the following
part of the tutorial, do not use the files opx.cif_pets and opx_dyn.cif_pets that you just
created, but use the files opx_start.cif_pets and opx_dyn_start.cif_pets provided with the
example files (Rename the files to opx.cif_pets and opx_dyn.cif_pets). Using your own
cif_pets files is possible, but your results may slightly differ from the results described in this
tutorial.
Start Jana2006
“File → Structure → New” opens a file manager
Left pane: locate the directory with input files
Right pane: double-click opx

526 Jana2006 Cookbook, version October 2017

3. Import Wizard
Select “Known diffractometer formats”; NEXT
Select “Pets electron diffractometer”
The file name automatically changes to opx.cif_pets, which is a file produced by PETS
containing all important information, including the list of intensities.
NEXT
Leave all settings unchanged; NEXT
Leave all settings unchanged; NEXT
The program reads 5232 reflections from the file
OK
For absorption correction select “None or done before importing”; NEXT
FINISH; OK; YES to accept the data set
You just read in the experimental parameters, cell parameters and the intensity list.

4. Symmetry Wizard
NEXT to close the information window and start the wizard
Symmetry Wizard can be started separately by “File → Reflection file → Make space group
test.
The default settings are prepared from x-ray diffraction. For electron diffraction the
tolerances should be increased so that Jana properly detect that the cell is orthorhombic.
Change the tolerances according to this screenshot:

Leave other settings default; NEXT; OK

[On the screen: Select Laue symmetry]
Select orthorhombic Laue symmetry; NEXT
The internal R-value of 23.68% is quite acceptable for EDT data.
[On the screen: Select cell centering]
The numbers in the table indicate that the only acceptable lattice type is primitive. In
general it is a good practice to check the centering in the reciprocal-space reconstructions.
Select centering P; NEXT
[On the screen: Select space group]
The window shows possible space groups. It can be quite tricky to determine the correct
space group form EDT data, as the systematic absences can be violated in electron
diffraction. You should use the reconstructions of the reciprocal-space layers to check, if you
can see some systematic absences. From the inspection of the section hk0.pgm (open the
file in ImageJ) the presence of the a-glide is quite obvious. The b-glide and c-glide conditions
in the files 0kl.pgm and h0l.pgm are violated by a couple of weak reflections, but still their
presence is very likely. Pbca is therefore a good first guess.

527 Jana2006 Cookbook, version October 2017

 Select Pbca; NEXT
Note that a wrong choice of the space group is not detrimental for the structure solution,
because we will use Superflip, which will make additional check for the most probable space
group.
Accept the space group in the standard setting; FINISH
Symmetry is saved in file opx.m50.

5. Creating refinement reflection file

In this step the program creates file opx.m90 containing the data set merged by symmetry
and with discarded forbidden reflections. M90 will be used for refinement.
Proceed with the wizard, leaving all settings default.

6. Structure Solution Wizard

OK to close the information window and start the wizard
[On the screen: window of Structure solution]
Structure solution wizard can be executed separately through “Run → Solution”.
Jana2006 does not contain solution procedures, it calls external programs. By default, Jana
uses Superflip (using charge flipping as the solution method). Superflip is distributed with
Jana2006.
In “Formula” textbox type the expected formula: Fe Mg Si2 O6
In “Formula units” textbox type 8
Click the checkbox next to “Use a specific random seed” and leave the value 111.
The specific random seed is used just to guarantee reproducibility of the results. For
standard work, random seed need not be defined.
Click the radio button “EDMA – fixed composition”
EDMA is an external program for interpretation of electron densities. It can take the
chemical formula and interpret the density using this formula.
Leave other settings default; OK
[On the screen: window of Superflip appears with iteration running. After ten runs, listing of
Superflip log file is displayed]
Superflip performed ten solution attempts. Check the space groups derived by Superflip from
each of the attempts. In eight out of ten cases the proposed space group is Pbca. It is thus
likely (though not guaranteed!) that Pbca is indeed the correct space group.
Press CLOSE to leave the listing but stay in the dialog “Structure solution
The result is a file opx.m40 with atoms of the structure model.

7. Verification of the structure solution

Click “Draw 3D map“
VESTA opens with a density shown as isosurfaces, and an atomic model overlaid. Rotate the
view to inspect the structure. Try to recognize the structural motifs. The main expected
motifs are SiO4 tetrahedra and (FeMg)O6 octahedra. The analysis will be simpler if you
display the atomic bonds:
Go to Edit → Bonds, and set up the options according to this screenshot:

528 Jana2006 Cookbook, version October 2017

 Viewing the structure solution in VESTA has the advantage that you can directly check, if the
raw density was correctly interpreted. In this case, the solution is very good and all atoms
are properly located. The only problem is that some cations have been assigned a wrong
atomic type. By inspecting the structure you should come to the conclusion that the
following two atoms need to be changed:
Mg1 -> Si
Si2 -> Mg or Fe – we will choose Mg for the moment
Quit VESTA (do not save changes)
Click “Accept last solution”

8. Change the atomic types

Double click icon Edit atoms
Double click atom Si1
Change its name to Mg2
Change its atomic type to Mg
OK
Repeat the procedure for the atom Mg1 and rename it to Si2
Close the dialog for editing atoms and confirm changes

9. Structure refinement
[On the screen: basic window of Jana]
Double click the icon Refine.
Refinement converges smoothly to an R value about 25%.

The octahedral positions are not exclusively occupied by iron or magnesium, but by both
elements. We have to refine the mixed occupancy in these sites.
[On the screen: basic window of Jana]
Double click icon Edit atoms
Right-click atom Fe1 → Split atomic position
Set up the dialog according to the following screen shot:

529 Jana2006 Cookbook, version October 2017

Apply
Repeat the procedure with Mg2 to generate atom Fe2 at the same position.

The procedure generated a pair of atoms with restricted ADP parameters and coordinates
and with a restricted sum of occupancies. The only small problem (although rather esthetic
than real) is that in one pair the Fe atom is the “Master” and Mg the “slave”, while for the
other pair it is vice versa. We will modify the restrictions to be both the same.
[On the screen: basic window of Jana]
Right-click Refine → Various → Restrictions
Load the restriction on atoms Mg2 and Fe2
Exchange the order of atoms in the restriction (type Fe 2 Mg2 in the line
“Atoms/molecules), so that Fe is the “master” atom. Rewrite the restriction. When finished,
the dialog should look like this:

OK; OK; Yes (do not start refinement yet!)

Now we have restricted the atomic pairs, but the occupancies are still fixed and would not
be refined. We have to allow their refinement explicitly:
Double click the icon “Edit atoms”
Double click atom Fe1
On the card “Edit” check the box next to the parameter “ai”; OK
Repeat for atom Fe2
OK to close the Edit atoms dialog

530 Jana2006 Cookbook, version October 2017

 Double click Refine to start refinement
Refinement converges to R-value of 24.25%. The refined atomic occupancies are:
Fe1: 0.44(5)
Fe2: 0.44(5)
Note that the equal occupancies and their e.s.d. is a pure coincidence and not a result of
some restriction or symmetry.

10. Set up dynamical refinement

At this moment there is very little we can do with the kinematical refinement. We have a
satisfactory refinement, the structure makes sense, the occupancies have been refined, but
their accuracy cannot be estimated from the kinematical refinement. Furthermore, some
ADPs are negative. We will now proceed to the structure refinement using dynamical
diffraction theory.
File → Structure → Save as
Save the structure as opx_dyn
Answer Yes to the question if you want to continue with the new structure
File → Reflection file → Import/modify the reflection file
Reimport; Browse for the file opx_dyn.cif_pets; NEXT; NEXT; NEXT; OK; NEXT; Finish; OK;
Yes
You have just replaced the previous reflection data with new data set suitable for dynamical
refinement.
Cancel to avoid the production of refinement reflection file
If we continued, we would have produced a symmetry-averaged file for kinematical
refinement. For dynamical refinement the intensities on each frame must be kept separate.
We will now enable the dynamical refinement option:
Go to “Parameters → Electron diffraction”
Check the checkbox “Use dynamical approach”
Proceed with creation of the reflection file, leave all settings default.

At this point we are ready to start dynamical calculations. Because the calculations are very
time consuming, we will not, in this example, use all 90 zones, but only nine selected zones.
Click “Select zones for refinement”; In the line below the table, set the following:

Click the button ; OK

Only zones 10, 20, 30, 40, 50, 60, 70, 80 and 90 are now selected for refinement.

The data collection was performed on one crystal and there is also no reason to expect that
the thickness varies for each zone. We will therefore restrict the thickness to the sam value
for all zones:
Click “Define zones of equal thicknesses; Select all; OK

Dyngo – the external executable that performs the heavy calculations – can run in a parallel
mode. In the field “Number of threads” you can decide, how many parallel threads Dyngo
may use. For the fastest performance set this number to the number of cores of your
processor. Important! Most Intel processors indicate twice as many threads available than

531 Jana2006 Cookbook, version October 2017

they have cores due to the hyper-threading technology. Do not set the number of threads to
the number of threads available on your processor, but to the number of ture distinct cores!

As a first step we have to check, if our model fits the data sufficiently well to start the
refinement. For that, we will run thickness optimization. This procedure is relatively fast, and
for each selected zone it calculates an R-value as a function of thickness. For good data set
and reasonable starting model, this curve should have a minimum at a meaningful
thickness, and all curves should indicate approximately the same thickness.
Check the checkbox next to “Thickness” and click
Jana calls Dyngo for each zone. When the calculation is finished, each selected zone has next
to the zone number in the dialog the best R-value, e.g.:

All zones’ R-values are lower than 20%, some of them even lower than 10%. This is a pretty
good start.
Click “Show thickness plots”
Now you can view the curves R-value vs. thickness for all zones. In most cases the curve has
a well-defined minimum, only in zones 30 and 70 the minimum is at the lowest thickness.
The best thickness varies in a relatively broad interval, but the minima are broad, and our
structure model is far from perfect at this stage, so this is not surprising.
Quit the viewer of the thickness curves

We want the thickess to be the same for all zones in the refinement. Now the thicknesses
are all different and we should reset them to a common most appropriate value. The best
practice seems to be a visual estimation of the “typical” thickness in the thickness plots. In
this case 500 angstroms seems to be a good guess. Do not worry about the inaccuracy of
such guess – the thickness is a very robust parameter and it will refine to the correct value
even from a quite distant start.
Click “Select zones for editing”; “Select all”; OK
Double click the yellow field EDThick and type “500” in the field
This procedure reset the thickness of all zones to 500 angstroms.
We have to recalculate the scale factors for each zone, because the scale changes with
changing thickness:
Click the button without any other checkbox checked (neither thickness nor
Orientation)
When the calculation finishes,click OK to close the electron diffraction dialog.

You can now start the structure refinement.

[On the screen: basic window of Jana]
Double click the icon Refine
After the first cycle the refinement R-value Robs is 13.13%. After 5 cycles the refinement
should converge to an R-value of 8.29%.

In the next step we will optimize the orientation of each zone. Due to the inaccuracy of
goniometer positioning and orientation matrix, the positional angles of each diffraction
pattern are not exactly correct, and a correction should be found.
Go to “Parameters → Electron diffraction”
Check the checkbox next to “Orientation”
532 Jana2006 Cookbook, version October 2017
Click the button
An optimization procedure starts that finds the best zone-axis orientation for each pattern.
Wait until the calculation finishes. Now you can now run a few more cycles of structure
refinement.
Click OK to close the form “Electron diffraction”
Double click the icon Refine
The refinement converges after 10 cycles to an R-value of 5.10%. Inspection of the structure
model shows that the occupancies of Fe1 and Fe2 refined to 0.33(3) and 0.06(2),
respectively. This is still relatively far from the values from x-ray refinement, which are
0.432(3) and 0.157(3). However, there are only 188 observed reflections in this limited
dataset for the refinement of 52 parameters. The data to parameter ratio is thus poor. If all
90 zones are included in the refinement and the whole procedure is repeated with them, the
final result is:
Robs=5.89%
occ(Fe1)=0.389(7)
occ(Fe2)=0.148(7)

533 Jana2006 Cookbook, version October 2017

534 Jana2006 Cookbook, version October 2017
Example 20.1: CeAsSe
Merohedric twin with extinction imported from SHELX to Jana2006.
Chemical formula: CeAsSe
Reference: Structure provided by Max Planck Institute for Chemical Physics of Solids
Input files: CeAsSe.hkl, CeAsSe.ins

1. Creating new jobname

Start Jana2006
“File → Structure → New” opens a file manager
Left pane: locate the directory with input files
Right pane: double-click CeAsCe

2. Importing structure from SHELX

[On the screen: Specify type of the file to be imported]
Select “Structure – from SHELX”; Next
In File manager double-click CeAsSe.ins
The program reads the structure, detects the presence of CeAsSe.hkl and starts import of
reflections
Go through the usual steps of the Import Wizard
Rint = 4.69%

3. Checking the twinning of the imported structure

Start EditM50
The twinning check box must be selected
Press “Matrices”
The twinning matrix should be 0 1 0 | -1 0 0 | 0 0 1
Leave EditM50
Start “Parameters → Twin fractions”
The parameter “twwol2” must be 0.46346
Activate refinement of “twvol2”
Save the changes and leave to the basic window

4. Refinement
Run refinement
It converges with R=3.2%
Open refinement listing, GoTo Statistics
Higher R value for the strongest reflections indicates that extinction correction is needed
Close the refinement listing
Start “Parameters → Extinction”
Activate Isotropic extinction with Gaussian model
Save changes and leave to the basic window
Repeat refinement
In converges with R=2.75%. The reason extinction has not been imported from the ins file is
that SHELX and Jana use different parameters for extinction correction

535 Jana2006 Cookbook, version October 2017

536 Jana2006 Cookbook, version October 2017
Example 20.2: HfPdGe
Six-fold twin with reflections in HKLF5 format, imported from SHELX to
Jana2006.

Chemical formula: HfPdGe

Reference: Structure provided by Max Planck Institute for Chemical Physics of Solids
Input files: HfPdGe.hkl, HfPdGe.ins

1. Creating new jobname

Start Jana2006
“File → Structure → New” opens a file manager
Left pane: locate the directory with input files
Right pane: double-click HfPdGe

2. Importing structure from SHELX

[On the screen: Specify type of the file to be imported]
Select “Structure – from SHELX”; Next
In File manager double-click HfPdGe.ins
The program reads the structure, detects the presence of HfPdGe.hkl, automatically
recognizes HKLF5 format and starts import of reflections
Go through the usual steps of the Import Wizard
Rint = ----

3. Checking the twinning of the imported structure

Start EditM50
The twinning check box must be selected, number of twin domains should be six
Press “Matrices”
The twinning matrices are all unit matrices because for HKLF5 data they have just a formal
meaning
Leave EditM50
Start Parameters → Twin fractions
The parameters “twwol2”, …. “twvol6” must contain the numbers from the BASF commans
of the ins file.
Activate refinement of all twin volume fractions
Save the changes and leave to the basic window

4. Refinement
Run refinement
It converges with R=4.1% for 60 parameters. There were very large changes of the twin
fractions during the refinement because probably the order of twin fractions in SHELX and
Jana differs.
Open refinement listing, GoTo Statistics
There is no indication of extinction and at the bottom of the listing no warnings.

537 Jana2006 Cookbook, version October 2017

538 Jana2006 Cookbook, version October 2017
References
List of references related to methods used in Jana2006, including external
programs for charge flipping and electron diffraction.

Revised: 10 May 2015

Petricek, V., Coppens, P. and Becker, P. (1985). Acta Cryst. A41, 478-483
Structure-analysis of displacively modulated molecular-crystals
Petricek, V. and Coppens, P. (1988). Acta Cryst A44, 235-239
Structure-analysis of modulated molecular-crystals .3. Scattering formalism and symmetry
considerations - extension to higher-dimensional space-groups
Petricek. V. (1989). Act aCryst. 45, 109-111
The possible translational parts of (3+d) superspace symmetry operations
Petricek, V., Maly, K., Coppens, P., Bu, X., Cisarova, I. and Frost-Jensen, A. (1991). Acta Cryst.
A47, 210-216
The description and analysis of composite crystals
Petricek, V., van der Lee, A. and Evain. M. (1995). Acta Cryst. A51, 529-535
On the use of crenel functions for occupationally modulated structures
Palatinus. L. (2007). Acta Cryst. A60, 127-237
Ab initio determination of incommensurately modulated structures by charge flipping in
superspace
Palatinus, L. and Chapuis, G. (2007). J. Appl. Cryst., 41, 786–790
SUPERFLIP - a computer program for the solution of crystal structures by charge flipping in
arbitrary dimensions
Baerlocher, C., McCusker, L. B., Palatinus, L. (2007). Z. Kristallogr. 222, 47-53
Charge flipping combined with histogram matching to solve complex crystal structures from
powder diffraction data
Leineweber, A. and Petříček, V. (2007). J. Appl. Cryst. 40, 1027-1034
Microstrain-like diffraction-line broadening as exhibited by incommensurate phases in powder
diffraction patterns
Petříček, V., Fuksa, J. and Dušek, M. (2010). Acta Cryst., A66, 649–655
Magnetic space and superspace groups, representation analysis: competing or friendly
concepts?
Palatinus, L., Fleischer, F. Pattison, P., Weber. T. and Steurer, W. (2011). Acta Cryst. A67, 9-20.
Ab initio reconstruction of difference densities by charge flipping
Perez-Mato, J.M., Ribeiro, J.L., Petříček, V. and Aroyo, M.I. (2012). J.Phys.: Condens. Matter, 24,
163201
Magnetic superspace groups and symmetry constraints in incommensurate magnetic phases

539 Jana2006 Cookbook, version October 2017

Palatinus, L., Jacob, D., Cuvillier, P., Klementová, M., Sinkler, W. and Marks, L.D. (2013). Acta
Cryst, A69, 171-188
Structure refinement from precession electron diffraction data
Palatinus L. (2013). Acta Cryst. B69, 1-16.
The charge-flipping algorithm in crystallography
Petříček, V., Dušek, M. and Palatinus, L. (2014). Z. Kristallogr., 229, 345–352
Jana2006: general features
Palatinus, L., Petříček, V. and Correa, C.A. (2015). Acta Cryst., A71, 235-244
Structure refinement using precession electron diffraction tomography and dynamical
diffraction: theory and implementation
Palatinus, L. (2015). Acta Cryst. B71, 125-126
Taking a closer look for a broader view: combining powder diffraction with electron
crystallography for a better understanding of modulated structures

540 Jana2006 Cookbook, version October 2017

Appendix 1.1: File format PRF
PRF file is used for powder refinement. It contains Bragg positions and profile
points.

Format of PRF is sometimes requested by users to obtain data points of powder

profiles for different purposes. For work with powders in Jana2006, knowledge of
the PRF format is not needed.
Revised: 10 May 2015

As Jana2006 allows refinement against more data sets, PRF file may consist of several data blocks:

Block1 begin
.........................
.........................
Block1 end
Block2 begin
.........................
.........................
Block2 end

In the case that there is just one data block, the labels are omitted.

Each block consists of three parts: header, Bragg positions and profile points:
2 0 0 2 3 3
1 -1 2 3. 1 10.9743 0.0237 0.0380 0.547364E+01
1 0 -2 3. 1 10.9743 0.0237 0.0380 0.547364E+01
1 -1 -4 3. 1 13.7339 0.0236 0.0414 0.698639E+01
1 0 4 3. 1 13.7339 0.0236 0.0414 0.698639E+01
0 0 -6 1. 1 14.2831 0.0236 0.0421 0.130568E+01
0 0 6 1. 1 14.2831 0.0236 0.0421 0.130568E+01
2 -1 0 6. 1 17.1683 0.0235 0.0456 0.668545E+01
2 -1 -3 6. 1 18.6097 0.0235 0.0474 0.768642E-01
2 -2 4 3. 1 22.0509 0.0234 0.0517 0.238082E+01
2 0 -4 3. 1 22.0509 0.0234 0.0517 0.238082E+01
2 -1 6 6. 1 22.4023 0.0234 0.0522 0.573109E+00
2 -1 -6 6. 1 22.4023 0.0234 0.0522 0.573109E+00
1 0 1 8. 2 25.3002 0.0232 0.0598 0.194769E+02
....................................
....................................
999
6.000000 0.209300E+04 0.249015E+04 0.457493E+02 6.023784 0 0.000000E+00 0.000000E+00 0.249015E+04
6.016766 0.209500E+04 0.247949E+04 0.457712E+02 6.040549 0 0.000000E+00 0.000000E+00 0.247949E+04
6.033532 0.217100E+04 0.246891E+04 0.465940E+02 6.057315 0 0.000000E+00 0.000000E+00 0.246891E+04
6.050298 0.206100E+04 0.245839E+04 0.453982E+02 6.074081 0 0.000000E+00 0.000000E+00 0.245839E+04
6.067064 0.204200E+04 0.244793E+04 0.451885E+02 6.090847 0 0.000000E+00 0.000000E+00 0.244793E+04
6.083830 0.217300E+04 0.243755E+04 0.466154E+02 6.107612 0 0.000000E+00 0.000000E+00 0.243755E+04
6.100596 0.210500E+04 0.242723E+04 0.458803E+02 6.124378 0 0.000000E+00 0.000000E+00 0.242723E+04
...............
...............
10.593884 0.102600E+04 0.986969E+03 0.320312E+02 10.617599 1 0.000000E+00 0.000000E+00 0.986969E+03
10.610650 0.108400E+04 0.101817E+04 0.329242E+02 10.634364 1 0.340796E+02 0.000000E+00 0.984355E+03
10.627417 0.991000E+03 0.101999E+04 0.314802E+02 10.651131 1 0.385393E+02 0.000000E+00 0.981752E+03
10.644182 0.105600E+04 0.102260E+04 0.324962E+02 10.667895 1 0.437751E+02 0.000000E+00 0.979160E+03
10.660948 0.104400E+04 0.102623E+04 0.323110E+02 10.684661 1 0.500343E+02 0.000000E+00 0.976579E+03
...............
...............
24.710857 0.361000E+03 0.316545E+03 0.190000E+02 24.734123 1 0.000000E+00 0.000000E+00 0.316545E+03
24.727623 0.347000E+03 0.316369E+03 0.186279E+02 24.750887 1 0.000000E+00 0.000000E+00 0.316369E+03
24.744389 0.341000E+03 0.353730E+03 0.184662E+02 24.767653 1 0.000000E+00 0.378276E+02 0.316194E+03
24.761154 0.353000E+03 0.356299E+03 0.187883E+02 24.784416 1 0.000000E+00 0.405929E+02 0.316019E+03
24.777920 0.385000E+03 0.359047E+03 0.196214E+02 24.801182 1 0.000000E+00 0.435381E+02 0.315845E+03
24.878517 0.375000E+03 0.383491E+03 0.193649E+02 24.901773 1 0.000000E+00 0.692188E+02 0.314806E+03
...............
...............
999.

541 Jana2006 Cookbook, version October 2017

Header
kType 2 for Jana2006
KADoublet 0 data for data measured with wave length
1 for Kα1 and Kα2
DataType 0 X-ray/neutron CW data
1 TOF data as a function of time
2 TOF data in d values
3 energy dispersive data
NPhases number of phases
(NDim(i),i=1,NPhases) number of indices used for each phases

Bragg positions
h,k,l,... n diffraction indices, where n = max(NDim(1), …, NDim(NPhases)).
For phases having the dimension smaller than n the indices are completed
by zeros.
Multiplicity multiplicity of the reflection following from the symmetry
KPhase defines the phase the reflection belongs to
The following four items are repeated twice if KADoublet=1. The first block corresponds to Kα1
TwoTheta/Time angular position/time (or for TOF d value/energy) of the peak as calculated
/d/energy from cell parameters and modulation vectors
Shift the angular shift of the peak position due to a zero shift, sycos and sysin
parameters (it need not be constant for all reflections). The actual of
reflections with respected to the profile should be calculated as:
[TwoTheta/Time/d/energy] -Shift
FWHM in the same units are previous values
Icalc(h,k,l) intensity as calculated from Rietveld (corrected for multiplicity and
preference orientation) or as extracted from the profile by le Bail fit.
The list of Bragg peaks ends with the integer 999 at the position of the h index.

Profile points
TwoTheta/Time/d/energy diffraction angle/time for TOF/d value/energy of the profile
point
Iobs observed intensity
Icalc calculated intensity
sig(Iobs) standard uncertainty of Iobs
TwoTheta-corr/Time-corr/ diffraction angle/time for TOF/d value/energy of the profile
d-corr/energy-corr point corrected for zero shift, sysin and sycos
Icalc calculated intensity
Skip flag 0 if the profile point was used in the refinement
1 if the profile point belongs to one of excluded region
Icalc(1) calculated intensity of the first phase
....
Icalc(NPhases) calculated intensity of the first phase
Ib background intensity
The list of peak points ends with the real number “999.”.

542 Jana2006 Cookbook, version October 2017

Appendix: Refinement
Refinement is the most important procedure of Jana2006. This appendix gives
theoretical background for refinement and explains some general technical
features of the user interface.

Revised: 19 November 2013

Refinement on F and on F2

The old programs used usually the refinement on F. This means that the minimized function has
the following form:
S F = ∑ wi Fi ,obs − Fi ,calc ( )2

i (1)
wi = 1 σ (Fi ,obs )
2
where . The refinement based on F2 uses an analogical formula:

S F 2 = ∑ wi (Fi ,2obs − Fi ,2calc )

2

i (2)
wi = 1 σ (F ). According to the error propagation theorem standard uncertainties
2 2
but now i ,obs

σ (F ) = 2 F σ ( F ) 2
are related by the equation which leads to:
i ,obs i ,obs i ,obs

SF2 = ∑
1
(F −F ) =∑
2 (F i ,obs + Fi ,calc ) (F
2

− Fi ,calc ) 2

(F ) (F )
2 2

σ σ
2 2 i ,obs i ,calc 2 2 i ,obs
4F 4F
(3)
i i ,obs i ,obs i i ,obs i ,obs

Fi ,obs ≈ Fi ,calc
In final stages of the refinement and therefore as an approximation we have:
2

SF2 ≈ ∑
4 Fi ,obs
(F − Fi ,calc ) = ∑ σ ( F1 ) ( F
2
− Fi ,calc )
2
= SF
i
2
4 Fi ,obs σ 2 Fi ,obs ( ) i ,obs
i
2
i ,obs
i ,obs

This means that the results for most cases should not give so different results.
However more precise analysis gives:

SF2 = ∑
(2 F i ,obs − Fi ,obs + Fi ,calc ) (F − F ) =
2
2

i 4 Fi ,2obsσ 2 Fi ,obs ( ) i ,obs i ,calc

) ( F − F ) + ∑ 4F )( F )
1 1
SF − ∑
3 4
− Fi ,calc
i Fi ,obs σ 2
(F i ,obs
i ,obs i ,calc
i
2
i ,obs
2
(
σ Fi ,obs
i ,obs

S
This means analogically that difference between the minimized functions F and S F is more
2

and more negligible when we approach the minimum.

The main advantage of refinement on F2 is that it uses as the weights as derived from
σ (Fi ,obs
2
) σ (Fi ,2obs ) = 2 Fi ,obs σ Fi ,obs ( ) leads to singularities for
experimental . The formula

543 Jana2006 Cookbook, version October 2017

Fi ,obs → 0
. Another problem is that intensities of some reflections can be negative due to
random errors and there is no way how to express such an experimental distribution when
using F instead of F2.
The above formulas can also be used to show why the R values base on F2 refinement are
higher and how we can compare them with those on F. Thus for the weighted R value we have:

∑ 4F
1
(F − Fi ,2calc )
2
∑ σ ( F ) (F
1
− Fi ,calc )
2
∑ σ ( F ) (F
1
− Fi ,calc )
2

(F )
2

σ
2 2 i , obs 2 i , obs 2 i , obs
i i , obs i , obs i i , obs i i , obs
RwF 2 = = =
1 1 1
∑ 4F
i
2
σ 2
(F ) Fi ,2obs ∑ 4F
i
2
σ 2
(F ) Fi ,4obs ∑ 4σ ( F ) F
i
2
2
i , obs
i , obs i , obs i , obs i , obs i , obs

∑ σ ( F ) (F − Fi ,calc )
1 2
2 i ,obs
i
RwF 2 ≈ 2 = 2 RwF
i ,obs

1
∑i σ 2 F Fi,2obs
( )
i ,obs

which means that the weighted R values based on F2 is about two times larger than that based
on F.
As obvious from the previous the weighted R values are closely related to the minimized
function and therefore they should reach their minimal values mutually.
On the other hand the non-weighted R values are defined:
∑F −F i ,obs i ,clac ∑ F −F 2
i ,obs
2
i ,calc
RF = i
RF 2 = i

∑F i
i ,obs ∑F i
2
i ,obs

and they need not to reach their minimal values at the same time as the minimized function.
RwF 2 RF 2 ≈ 2 RF
Similarly as for and RwF we can get
RwF 2
In the output of Jana2006 you can find either RwF or depending on the type of refinement.
As a non-weighted R value we use always RF .

544 Jana2006 Cookbook, version October 2017

Volume fractions

The volume fraction vi is defined as a ratio of the individual domain Vi volumes to the sample
volume V :
V
vi = i
V
In the case of the twinned crystal with complete overlaps of diffraction spots (merohedric twins
or pseudo-merohedric twins with lattice fulfilling higher symmetry within experimental errors)
this volume fractions can directly be used to express a “combined structure factor” F
2
(h )
N
F 2 (hTi )
F 2
(h ) = ∑ vi 2
i =1 Vcell
where F (h ) is the structure factor, Vcell is the cell volume and Ti are the twinning matrices
and N is the number of diffractionally independent twin domains.
The formula can be generalized in the cace for general pseudo-merohedric twin:

(h ) = ∑ vi pi F (hTi)
N 2
F 2
2
i =1 Vcell
where the factor pi is equal 1 if the diffraction spot of i-th domain, having indices hTi , is
contributing to the spot localized at the point h and otherwise it is equal zero. Note that in this
case the diffraction vectors hTi need not have integer components.
Another generalization can be made for multiphase samples. Then the previous formula is
changed to:
N F j2 (hTi )
F 2
(h ) = ∑ vi 2
i =1 Vcell ,j
where the index j defines the j-th phase domain.
Similar formulas are used in the Rietveld refinement. Then a contribution to the profile point is:
N F j2 (h i )
ycalc (2ϑ ) = Lp (ϑ )R(ϑ )∑ vi Ai (ϑ ).∑ Pi (h i )mi (h i ) 2
Ω i (h i ,ϑ )
i =1 hi Vcell ,i

Lp (ϑ ) Lorentz-polarization factor
R (ϑ ) roughness correction
Ai (ϑ ) absorption factor
Pi (h i ) preference orientation function
mi (h i ) multiplicity factor
Ωi (h i ,ϑ ) profile function
The volume fractions can be transformed to the relative phase amounts in mass by the formula:

vi ρ i
ri= N

∑v ρ
i =1
i i

where ρi are the density of the i-th pahse.

545 Jana2006 Cookbook, version October 2017
Appendix: Refinement keys
Restrictions.

Revised: 19 November 2013

Automatic/manual handling of refinement keys

All refinable parameters are divided into two groups: safe and problematic ones. The safe
parameters are those which can be usually refined without numerical problems: scale factors
and extinction parameters; parameters describing position, ADP and their modulation for
individual atoms as well as for molecules; background parameters of the powder profile. The
problematic parameters are twin/phase volume fractions, atomic or molecular site occupancies,
parameters of crenel modulation functions, parameters of multipole refinement and powder
profile parameters for individual Bragg reflections.
There are two ways how to handle refinement keys in Jana2006. The default one is the automatic
mode in which refinement keys of all “save” structural parameters are set “on” before the
refinement procedure. The refinement keys can also be handled in the manual mode and then
we have to specify explicitly all refinement keys for all parameters.
Selection of the refinement mode can be made in “Refine commands”:

The both modes can be combined with the so call “fixed” commands which can fix different
groups of parameters. The following fixed command fixes for example coordinates of all atoms:

546 Jana2006 Cookbook, version October 2017

The choice “Individual” and “Set individual” allow fixing and setting individual parameters. Here
we can use wild characters to define groups of parameters and/or groups of atoms to which
the fixed command is applied. We can for example write:

which means that cos modulation wave of all first harmonics for atoms having name started with
the character “C” with be set to zero and kept unrefined. For this is important that every
parameter is Jana2006 has its own name which is used in all output and input procedures.

Another example can fix all multipole parameters

547 Jana2006 Cookbook, version October 2017

The fixed commands are also active in the manual mode and then they combined with user’s
selection of refinement keys.

548 Jana2006 Cookbook, version October 2017

Appendix: Modulation
functions
Modulation functions.

Revised: 28 December 2013

Abstract

A full solution and refinement of modulated structures, having strong occupational modulation described
by crenel functions, is possible only if modulations of other parameters are described with respect to a set
of functions orthogonalized in the crenel interval. It is shown how such modulation can be recognized
from so called “de Wolff’s sections”. Moreover, two new sets of functions (Legendre polynomials and x-
harmonics) were introduced and tested on simulated and real structures with a new version of program
Jana2006.

1. Introduction

The superspace theory as developed by de Wollf et al., (1981) proved to be the best way how to generalize
methods for solution and refinement of regular structures to modulated and composite ones. Any structural
parameter (occupancy, atomic coordinates and ADP) can be expressed as a periodic modulation function
of the internal coordinate x4 :

p ( x4 ) = p ( x4 + 1) (1)

The actual value of the structural parameter p of the atoms located at the position r0 + n , where n defines

the position of the actual cell, is calculated from the equation (1) with x4 = q ⋅ (r0 + n ) + t . Here q stands
for the modulation vector and t defines a phase factor defining a "first" cell.

The fact that the modulation function is periodic allows its expansion into a Fourier series:

p(x4 ) = p0 + ∑ ( p sn sin 2πnx4 + p sn cos 2πnx4 ) (2)

n

549 Jana2006 Cookbook, version October 2017

Thus the modulated structure is fully described if the set of coefficients p0 , p sn , pcn is determined for each
structural parameter. Depending on a particular structure the number of terms used in the equation (2) can
be different but usually the number of harmonics is not higher that four.

But as has been already shown (Petříček et al., 1990 and Petříček et al., 1995) there are cases where the
number of harmonics, necessary to describe a modulation function, is very large. This is true especially of
modulation functions exhibiting some discontinuity - such as positional saw-tooth function and
occupational crenel function.

It turns out that with growing number of successfully solved structures the number of cases, having some
type of discontinuity in modulation functions, increases. These functions are also often used to describe by
a unified way families of related structure ( e.g. Evain et al., 1998; Peréz-Mato et al., 1999). Therefore in
the first part of the paper it is shown how such discontinuities modulations can be correctly recognized by
analyzing of de Wolff’s sections through periodic (3+1) dimensional Fourier maps.

The basic problem in application of the crenel and saw-tooth functions is that harmonic functions (2)
defined in the interval (0,1) are not anymore orthogonal when used in the shorter interval reduced by the
crenel function. We shall call it in following text the crenel interval. This leads to strong correlations
between coefficients of positional or ADP modulations. On the other hand harmonic functions, defined in
the full interval, allow an expansion of the structure factors into a series of products of Bessel functions
(Peréz-Mato, et al.,1986 and Petříček & Coppens, 1988) which was the only method used in the old version
of the Jana program. For this reason a combined method of selection of harmonics, defined in the full
interval, and their re-orthogonalization into the crenel interval has been developed (Petříček et al., 1995).
This method has been successfully applied by different authors to various modulated structures. The main
disadvantage of this method is that depending on the crenel interval the set of used functions (linear
combinations of basic harmonics) are generally different and therefore they must be always listed together
with set of refined coefficients to allow a full reproduction of refined modulation curves.

As the Bessel expansion method was much less efficient for more complicated modulated structures than
the integral method introduced in the program REMOS by Yamamoto, (1982) the newer versions of Jana
program are offering both methods for calculation of structure factors. This opened possibility to reconsider
way how to define modulation function within a crenel interval. In the section 3 two new orthogonal sets
of functions, defined directly in the crenel interval, are introduced and tested with simulated structures.

2. Interpretation of Fourier maps

As it has been presented in the old paper by de Wolff (1974), the generalized (3+1) dimensional maps based
on the already know model can be used to prove that the solution is correct or to find more details about
modulation functions. Similarly like for regular structure we can use Fourier, difference Fourier or Patterson
maps. Probably the most commonly used are 2d sections (x1-x4, x2-x4 or x3-x4) of (3+1) dimensional

550 Jana2006 Cookbook, version October 2017

electron or nuclear density in vicinity of the selected atom. These sections, usually called de Wolff's
sections, are used also to detect possible discontinuities in modulations.

However there are some effects which can flutter of such maps in so strong way that a wrong conclusion
can be drawn. In many cases a positional modulation affects more than one individual coordinate and the
modulation looks like for example helically or elliptically modulated distortions. Then the simple de Wolff's
sections can look like we can see in Fig. 1a. The fact that the atom is modulated in two directions leads to
artificial effects which could make a false impression of discontinuity in modulation. This can be solved
by calculating sections in other directions within some vicinity of the selected atom and projecting it (see
Fig 1b).

Another problem is connected with resolution caused by a termination effect in the Fourier summation. For
modulated structures the crucial point is if all detectable satellites were measured and used in the Fourier
calculation. This can be demonstrated by a simulated structure in which the atom has a strong harmonic
positional modulation along c axis with amplitude about 1 Å. Then the de Wolff's section calculated with
only first order satellites gives an impression that the atom is predominantly located in two distinct positions
as shown in Fig 2a. However the Fourier maps made with satellites up to fifth order (Fig 2b) shows clearly
that the modulation has purely harmonic character.

A possible smearing of a Fourier map can be estimated from so called shape function which is a Fourier

summation: S (r ) = ∑ A(H ) exp(− 2π r ⋅ H ) with A(H ) = 1 for all measured diffractions. For all others
H

diffraction points A(H ) = 0 . The calculated map is a convolution of an ideal map with the shape function.
The more indices are measured the more δ-like character the function S (r ) has and the more truthful the
map is. Fig 2c and 2d show the shape function for two simulated cases mentioned above. A strong
prolongation of the shape function along x4 axis, for the case in which only main reflections and first order
satellites are used, leads to severe smearing as visible in resulting map (Fig 2a). This is considerably
changing the character of the modulation curve in points where derivation of the ideal modulation curve
with respect to x4 is large.

On the other hand the refinement process is much less sensitive to the maximal order of used satellites. This
can be demonstrated by a simulated refinement against a data set made only from main reflections and first
order satellites. Then the refinement of a crenel-split model, as can be concluded from a Fourier maps based
on the same set reflections (see Fig 2b), gives a very bad fit. The reason for this disagreement can be
understood from differences between contributions to structure factors for a crenel-split model and the
model based on a strong positional modulation. In the simulation we used the modulation vector of form

q = γa*3 and then for the diffraction vector:

H = ha1* + ka*2 + (l + γm )a*3 = H 1a1* + H 2a*2 + H 3a*3

551 Jana2006 Cookbook, version October 2017
the contribution to the structure factor for the crenel-split model is (e.g. Petříček et al. (1995)):

sin πm∆
F (h, k , l , m ) ~ exp(2πiH ⋅ r0 ) sin[2π ( x3b − x3a )H 3 ] (3a)
πm

Where r0 is an average position of the split atoms, x3a and x3b actual z-coordinates of both atomic positions
and ∆ is crenel length, here equal to 1/2.
On the other hand the positional modulation leads to the following structure factor:

Fν (h, k , l , m ) ~ exp(2πiH ⋅r0 )J −m (2πu z H 3 ) (3b)

The fig. 3 shows that these two expressions, as a function of (x3b − x3a )H 3 and u3 H 3 , have a quite
different form even for first order satellites.
The refinement of the crenel-split model is to be combined with a positional modulation described by a set
orthogonalized modulation function as will be described later. Then one modulation wave is sufficient to
get a perfect fit and the modulation curve, as demostrated on the fig 4, which is almost identical with the
original harmonic wave.

From what was demonstrated in this chapter we can see that the de Wollf's sections can help to recognize
possible discontinuities in the modulation curves and get more realistic description of the modulated
structures. Such an analysis of (3+1) dimensional density maps can be also used to extract a correct solution
from maps supplied by Superflip (Palatinus & Chapuis, 2007).

3. Orthogonalized set functions for modulation of positional parameters and ADP

As it was mentioned in the introduction the crenel occupancy modulation leads to severe numerical
problems when combined with modulations of other parameters (position and ADP) if the standard

description with harmonic functions (2), defined in the interval 0,1 , are used. This problem has been
solved by Petříček et al. (1995) by an orthogonalization method combined with a selection of harmonics.
Here we would like to repeat the main ideas from this paper and give two more alternative methods for
description of modulations within the crenel interval. Finally all three methods are compared on a simulated
data set.

3.1. Harmonics defined in <0,1> interval orthogonized to the crenel interval

The main advantage of this method is that harmonic modulation functions

{g i (x ), i = 1,2n + 1} = {1, sin (2πx ), cos(2πx ),sin (2πnx ), cos(2πnx )} for atomic positions and harmonic
ADP's, defined in the interval 0,1 , leads to analytical expressions for a structure factor based on sum of
products of Bessel or modified Bessel functions which is true even for cases when such modulations are

552 Jana2006 Cookbook, version October 2017

combined with a crenel occupancy modulation. This was the main reason why the orthogonalization method
was preferred to some orthogonal set of functions defined directly in the crenel interval

x40 − ∆ 2, x40 + ∆ 2 .

The Gram-Schmidt orthogonalizion procedure is used to construct a new set of functions {γ i (x )}, already
orthogonalized in the crenel interval, as a linear combination of the original harmonic functions to

{g i (x ), i = 1,2n + 1}:
2 n +1
γ j (x ) = ∑ T ji g i (x ) (4)
i =1

The Gram-Schmidt orthogonalizion procedure leads to the lower triangular matrix T which is generally
different for different crenel intervals. This means for each particular used crenel function the set of
coefficients forming the matrix T must be presented to allow full reconstruction of the modulated
functions.

Another point is that for numerical reasons the othogonalization method is to be combined with a selection

procedure in which each new function say g j (x ) from the sequence {g i (x ), i = 1,2n + 1} is first tested if it

makes an angle α j to already chosen subset of functions larger than a chosen limit. For practical application

the acceptance condition is expressed as:

ε i = cosα i < λ (5)

As already mentioned in the original paper this leads to some incompleteness of the selected set of functions
but experience with application of this method to various compounds approved that our default value
λ = 0.95 is reasonable.

3.2. Orthogonal set of functions defined in the crenel interval

As already mentioned in the introduction, the analytical method for calculation of structure factors based
on Bessel functions was becoming less and less effective for complicated modulated structures and
integration method (Yamamoto, 1982) is now used almost exclusively. This called for reconsideration of
used set of orthogonalized sets defined directly in the crenel interval. Such function has two main
advantages to the original orthogonalized set. These sets are clearly and uniquely defined for any crenel
interval and therefore they have always the same mathematical form. Moreover, they form mathematically
complete sets of functions.

553 Jana2006 Cookbook, version October 2017

As a first candidate is a set polynomials {1, x, x , x ,} defined in the interval
2 3
− 1,1 . But as these

functions are not mutually orthogonalized again the Gram-Schmidt orthogonalizion procedure has to be
used. But this leads to a standard set of function depending on definition of the scalar product:

(g , g ) = ∫ w(x )g (x )g (x )dx
1

i j i j
−1

For different choice of the weight function w( x ) we can get e.g. Legendre, Chebyshev, Laguerre or Hermite
polynomials (Abramowitz & Stegun, 1972). The most natural choice for incommensurately structures is
the unit weight as all points in the interval where atom is present have the same probability. This condition
leads to the Legendre polynomials. For comparison of modulation amplitudes it important to take into
account that Legendre polynomials are not normalized but that the following equation holds:

( fn , fm ) = 2
δ nm
2+n

The crenel interval x4 − ∆ 2, x4 + ∆ 2 is transformed to

0 0
− 1,1 by transformation:

(
x = 2 x4 − x40 ∆ )
Another candidate for orthogonalized set of functions in the crenel interval is the set of harmonic functions
{1, sin (πx ), cos(πx ), sin (2πx ), cos(2πx ),sin (nπx ), cos(nπx )} . But as all these functions give the same
value at points -1 and +1 they cannot fully describe all possible modulations within the crenel interval as
generally the modulations need not be identical in the first and the last point of the crenel interval. For this
reason the set of functions is to be completed with a function which would correct for this. The simplest
way is to add to this sequence a linear term:

{1, x, cos(πx ), sin (πx ), cos(2πx ), sin (2πx ),, cos(nπx ), sin (nπx )}

Then in the sequence of functions all functions with odd order are even and vice versa. While the original
set of harmonic functions in the crenel interval is mutually orthonormal, the new set of functions is not as
the following formula holds:

1
(x, sin nπx ) = ∫ x sin (nπ )dx = (− 1)n+1 2
(6)
−1
nπ

On the other hand all other mixed scalar products are equal zero. But this sill means that the Gram-Schmidt
orthogonalizion procedure is to be used to get fully orthonormalized set. In the Fig.5 the orthonormalized
functions of order from 2 till 6 are visualized. The set of coefficients for first 20 functions and analytical
expressions as derived by the program Maple 12 are supplied as supplementary materials. To our

554 Jana2006 Cookbook, version October 2017

knowledge this set of orthogonalized functions has no special name and we shall call it in the following
text x-harmonics.

3.3. Comparison of three different orthogonal sets of function

The three sets functions - harmonics in 0,1 orthogonalized to the crenel interval (in the following shortly

"Ortho-harmonics"), Legendre polynomials and x-harmonics can give generally a different fit when used
to describe modulations in the crystal. The advantage of the Legendre polynomials and x-harmonics is that
they make a complete set of functions and therefore they can describe any modulation function if a sufficient
number of terms is used. In order to demonstrate different situations two simulated structures with structural
parameters as summarised in the Table 1 have been used. In the simulation procedure noise from a random-
number generator was applied to ensure conditions similar to real experiment.

Both cases exhibit very strong positional modulation within the crenel interval and they are very far from
what is usual for real structures. However such tests can show differences between using of different set of
functions. The results of three refinements, in which the same number of positional parameters was used as

for simulations, are summarized in the Table 2. R values for the case x40 = 0.2 is almost the same for all
sets but for the second case the x-harmonics makes considerably better fits.

A comparison how refined modulation curves fit to those from the simulated model are shown in the Fig 6

and 7 for the case x40 = 0.2 and x40 = 0.6 , respectively. For refined R values and graphic curves it is
obvious that the x-harmonics gives the best fit. However after making more tests, also with some already
refined and published structures, it turned out that differences are not so pronounced as for the selected
simulated structures and that all three methods could give similar results.

4. Conclusions

The original method for refining of modulations of structural parameters combined with crenel occupancy
modulations used the “ortho-harmonics”, as was described by Petříček et al (1995), have two main
disadvantages to the newly introduced orthogonalized sets of functions. Their coefficients are to be always
updated when the crenel interval is changed. This is especially crucial during the refinement process. The
second disadvantage is that modulations functions can be reproduced only if together with refined
amplitudes a set of actual orthogonalization coefficients is supplied.

The newly introduced sets of functions, Legendre polynomials and x-harmonics, are defined in the crenel
interval in a unique mathematical form and therefore they make the refinement smoother. While the
Legendre polynomials are standard functions x-harmonics, to our knowledge, were not yet used and
tabulated. For this reason all details how to use them are given as supplementary materials. From made
tests we found that the refinement with x-harmonics gives usually better fit than compared to Legendre
polynomials and that the normalized modulated amplitudes are smaller.

555 Jana2006 Cookbook, version October 2017

The latest version of the program Jana2006 allows application of both new sets of orthogonalized functions.
Moreover, the old method based on used the “ortho-harmonics” has been considerably improved and now
the necessary updating of orthogonalization coefficients is made automatically during the refinement
process.

Acknowledgements The work has been supported from Praemium Academiae of the Czech Academy
of Sciences and by the institutional research plan No. AVOZ10100521 of the Institute of Physics. One of
the authors (J.F.) acknowledges a financial support of the Grant Agency of the Academy of Sciences of the
Czech Republic under the grant reg. no. IAA100100918. One of the authors (V.E.) acknowledges a
financial support of Ministry of Education, Youth and Sports of the Czech Republic under grant reg. no.
MSM6046137302).

4.1. References

Abramowitz, M., & Stegun, I. A. (1972). Handbook of Mathematical Functions.

de Wollf, P.M., Janssen T. & Janner, A. (1981). Acta Cryst., A37, 625-636.
Evain, M., Boucher, F., Gourdon, O., Petříček, V., Dušek M. & Bezdíčka P. (1998). Chem. Matter., 14,
3068-3076.
Palatinus, L. & Chapuis, G. (2007). J.Appl.Cryst., 40, 786-790.
Peréz-Mato, J. M., Madariaga, G. & Tello, M.J. (1986). J.Phys.C, 19, 2613-2622.
Peréz-Mato, J.M., Zakhour-Nakhl, M. Weill, F. & Darriet, J. (1999). J.Mater.Chem., 9, 2795-2807.
Petříček, V. & Coppens, P. (1988). Acta Cryst., A44, 235-239.
Petříček, V., Dušek M. & Palatinus, L. (2006). Jana2006. The crystallographic computing system.
Petříček, V., Gao, Y., Lee, P. & Coppens P. (1990). Phys.Rev.B, 42, 387-397.
Petříček, V., Lee, A. van der & Evain, M. (1995). Acta Cryst., A51, 529-535.
Wolff de, P. M. (1974). Acta Cryst., A30, 777-785.
Yamamoto, A. (1982). Acta Cryst., A38, 87-92.

556 Jana2006 Cookbook, version October 2017

557 Jana2006 Cookbook, version October 2017
Table 1 Basic data for simulated structures

Formula: Ta
Basic unit cell: a = 3.018, b = 4.021, c = 3.815 Å, α = 80, ß = 75, γ = 80
Modulation vector: q = 0.3412a* + 0.4091b* + 0.2752c*
Superspace group: P 1 (αβγ )0
hklm range: all main reflections and satellites up to fifth order ( sin θ λ ≤ 1 Å-1)

Atomic position and amplitudes of the displacive modulation function

n Ax ,s ,n Ay , s , n Az ,s ,n Ax ,c ,n Ay ,c ,n Az ,c ,n
0 0.2500 0.2500 0.2500
1 0.0500 -0.0500 0.0500 -0.0250 -0.0350 0.0350
2 0.0250 0.0100 -0.0150 0.0500 0.0500 -0.0250
3 0.0350 0.0150 0.0300 0.0250 0.0150 0.000

Crenel function parameters: case#1 ∆=0.6 x40 = 0.2

case#2 ∆=0.6 x40 = 0.6
Harmonic ADP parameters

U 11 U 22 U 33 U 12 U 13 U 23
0.0150 0.0150 0.0200 0.0050 -0.0050 0.0050

Table 2 Results of refinements for two cases

Legendre
Ortho-harmonics x-harmonics
polynomial
Case RF RwF RF RwF RF RwF
x40 = 0.2 0.0703 0.0889 0.0752 0.0939 0.0675 0.0810

x40 = 0.6 0.1041 0.1289 0.0502 0.0625 0.0299 0.0341

558 Jana2006 Cookbook, version October 2017

Figure 1

2
0
2
0
x
0
x
x
0
0
,
x

x
4
x
4

6
6

2
2

0
8
0
8

0
4
0
4

0
0
0
0

-
0
2
0
-
0
2
0
0
0
0
x
x

(a) (b)

(a) x1-x4 de Wolff's section through the atom having helical distortion in the (x1, x2) plane. The electron
density shows an artificial discontinuity.

(b) x1-x4 de Wolff's section through the atom having helical distortion in the (x1, x2) plane with projected
section in x2-x4. The artificial discontinuity disappears.

559 Jana2006 Cookbook, version October 2017

Figure 2

1
.
0
1
.
0

x
4
x
4
x
0
0
0
0
,
x
2
0
0
0
0
x
0
0
0
0
,
x
2
0
0
0
0

0
.
8
0
.
8

0
.
6
0
.
6

0
.
4
0
.
4

0
.
2
0
.
2

0
.
0
0
.
0

-
0
.
2
0
-
0
.
2
0

0
0
0
0
0
0

x
x

(a) (b)

x
4
x
4

0
.
4
0
.
4
x
0
0
0
0
,
x
2
0
0
0
0x
0
0
0
0
,
x
2
0
0
0
0

0
.
2
0
.
2

0
.
0
0
.
0

-
0
.
2
-
0
.
2

-
0
.
4
-
0
.
4

-
0
.
2
0
-
0
.
2
0

0
0
0
0
0
0

x
x

(c) (d)

de Wolff's section of the atom having strong (1 Å) positional modulation.

(a) Fourier deWolff’s sections made from main and first order satellites

(b) Fourier deWolff’s sections made from main and satellites up to fifth satellite order

(c) Shape function made from main and first order satellites

(d) Shape function made from main and satellites up to fifth satellite order

560 Jana2006 Cookbook, version October 2017

Figure 3

Comparison of contribution to the structure factor for the split and harmonic model

561 Jana2006 Cookbook, version October 2017

Figure 4

1
.
0

x
4
x
0
0
0
0
,
x
2
0
0
0
0

0
.
8

0
.
6

0
.
4

0
.
2

0
.
0

-
0
.
2
0

0
0
0

0
0

Modulation curves refined with the split atom model described by crenel functions combined with
positional modulation.

562 Jana2006 Cookbook, version October 2017

Figure 5

Orthogonalized x-harmonics for n=2, 3, 4, 5 and 6.

563 Jana2006 Cookbook, version October 2017

Figure 6 Positional parameters x,y and z as refined for the case x40 = 0.2 . The red curve shows the
refined modulation curve while the black line shows the modulated wave as used in the simulation
procedure.
565 Jana2006 Cookbook, version October 2013
 (a) Ortho-harmonics (b) Legendre polynomials
(c) x-harmonics

Figure 7 Positional parameters x,y and z as refined for the case x40 = 0.6 . The red curve shows the
refined modulation curve while the black line shows the modulated wave as used in the simulation
procedure.

566 Jana2006 Cookbook, version October 2013

567 Jana2006 Cookbook, version October 2013
 (a) Ortho-harmonics (b) Legendre polynomials
(c) x-harmonics

568 Jana2006 Cookbook, version October 2013

Appendix: Powder parameters
Modulation functions.

Revised: 28 December 2013

Dear Pavel,

I wish you all the best in 2014. I am very sorry for so long silence but
during the Christmas time I was just relaxing.

Would you be so kind to give me a piece of advice for SizeL, StrainL, SizeG,
StrainG values handling? The question, namely, is the units of these values:
is it angstroms for Size values and dimensionless epsilon for Starin values,
or any coefficients are used (as for LX,LY and GV,GU,GW,GP in simple PV
model)?

In the case that the fundamental approach is used the number of profile
parametrs is reduced to (LX, LY, GU and GP):

H(Lorentzian) = LX/cos(theta) + LY*tan(theta)

H^2(Gaussian) = GP/cos(theta)^2 + GU*tan(theta)^2

But instead of refining LX and GP we are using parametrs SizeL and SizeG.
They are related to LX and GP by equations:

SizeL = 180*lambda(in nanometers)/(pi*LX)

SizeG = 180*lambda(in nanometers)/(pi*sqrt(GP))

Than these values are related to a means crystallize size in nanometers.

The strain parameters StrainL and StrainG are expressed in radians.

Enclosed you will my presentation about the powder option in Jana.

With best regards,

Vaclav

569 Jana2006 Cookbook, version October 2013

Dear Petricek, Dusek and Palatinus

In the paper by Leineweber and Mittemeijer (J.Appl.Cryst. (2004), 37,

123-135) they write that the anisotropic broadening parameters S_HKL in
JANA differ from the definition by Stephens (1999), e.g. that
S_220(stephens)=6*S_220(jana)

Is this still the case? Or is jana now in accordance with Stephens?

Sincerely
Sebastian

E-mail: [email protected]

Dear Sebastien,

Yes, you are right the definition in Jana is different than originally
introduced by Stephens. The reason for this is that in Jana we are using
the
procedure for applying symmetry restrictions of symmetrical 4th order
tensor
as used also for anharmonic ADP's. So I am not planning to modify it.

With best wishes,

Vaclav

Dear dr. In-Hwan Oh,

Thank you for your letter and intent in Jana2006. The asymmetry
correction named in Jana as "correction by divergence" is based on the
theoretical paper by Finger, Cox and Jephcoat attached to this letter.
The two parameters labeled in Jana as "H/L" and "S/L" are based on the
symbols used in that paper - 2H is the detector opening and 2S is the
sample height. Both are divided by "L" which is the sample to detector
distance.

Unfortunately these two parameters are usually strongly correlated and

this was the reason why we recommend users to fix these parameters by an
additional equation based on some estimation of real values S/L and H/L.
In Jana the equation can have the following form:

equation : H/L=S/L

570 Jana2006 Cookbook, version October 2013

and then only the S/L parameter is refined.

Recently Jim Hester published a paper in which he gives some improved

analytical formulas and he recommends to use:

(H-S)/L "in Jana "HmS/L and H+S/L in Jana "HpS/L". Than the condition
HmS/L=0 means H/L=S/L.

With best regards,

Vaclav

571 Jana2006 Cookbook, version October 2013